make the default JOB_EVENT_BUFFER_SECONDS 1 seconds (#14335)

Dispatcher refactoring to get pg_notify publish payload
  as separate method

Refactor periodic module under dispatcher entirely
  Use real numbers for schedule reference time
  Run based on due_to_run method

Review comments about naming and code comments

.github/ISSUE_TEMPLATE/bug_report.yml

@@ -44,6 +44,7 @@ body:
       label: Select the relevant components
       options:
         - label: UI
+        - label: UI (tech preview)
         - label: API
         - label: Docs
         - label: Collection

.github/dependabot.yml

@@ -1,19 +0,0 @@
-version: 2
-updates:
-  - package-ecosystem: "npm"
-    directory: "/awx/ui"
-    schedule:
-      interval: "monthly"
-    open-pull-requests-limit: 5
-    allow:
-      - dependency-type: "production"
-    reviewers:
-      - "AlexSCorey"
-      - "keithjgrant"
-      - "kialam"
-      - "mabashian"
-      - "marshmalien"
-    labels:
-      - "component:ui"
-      - "dependencies"
-    target-branch: "devel"

.github/issue_labeler.yml

@@ -6,6 +6,8 @@ needs_triage:
   - "Feature Summary"
 "component:ui":
   - "\\[X\\] UI"
+"component:ui_next":
+  - "\\[X\\] UI \\(tech preview\\)"
 "component:api":
   - "\\[X\\] API"
 "component:docs":

.github/pr_labeler.yml

@@ -15,5 +15,5 @@
 
 "dependencies":
   - any: ["awx/ui/package.json"]
-  - any: ["awx/requirements/*.txt"]
-  - any: ["awx/requirements/requirements.in"]
+  - any: ["requirements/*.txt"]
+  - any: ["requirements/requirements.in"]

.github/workflows/devel_images.yml

@@ -48,8 +48,11 @@ jobs:
           DEV_DOCKER_TAG_BASE=ghcr.io/${OWNER_LC} COMPOSE_TAG=${GITHUB_REF##*/} make awx-kube-dev-build
           DEV_DOCKER_TAG_BASE=ghcr.io/${OWNER_LC} COMPOSE_TAG=${GITHUB_REF##*/} make awx-kube-build
 
-      - name: Push image
+      - name: Push development images
         run: |
           docker push ghcr.io/${OWNER_LC}/awx_devel:${GITHUB_REF##*/}
           docker push ghcr.io/${OWNER_LC}/awx_kube_devel:${GITHUB_REF##*/}
-          docker push ghcr.io/${OWNER_LC}/awx:${GITHUB_REF##*/}
+
+      - name: Push AWX k8s image, only for upstream and feature branches
+        run: docker push ghcr.io/${OWNER_LC}/awx:${GITHUB_REF##*/}
+        if: endsWith(github.repository, '/awx')

.github/workflows/label_issue.yml

@@ -6,9 +6,9 @@ on:
       - opened
       - reopened
 
-    permissions:
-      contents: read # to fetch code
-      issues: write # to label issues
+permissions:
+  contents: write # to fetch code
+  issues: write # to label issues
 
 jobs:
   triage:
@@ -17,7 +17,7 @@ jobs:
 
     steps:
       - name: Label Issue
-        uses: github/issue-labeler@v2.4.1
+        uses: github/issue-labeler@v3.1
         with:
           repo-token: "${{ secrets.GITHUB_TOKEN }}"
           not-before: 2021-12-07T07:00:00Z

.github/workflows/label_pr.yml

@@ -8,7 +8,7 @@ on:
       - synchronize
 
 permissions:
-  contents: read # to determine modified files (actions/labeler)
+  contents: write # to determine modified files (actions/labeler)
   pull-requests: write # to add labels to PRs (actions/labeler)
 
 jobs:

.github/workflows/pr_body_check_jira.yml

@@ -0,0 +1,35 @@
+---
+name: Check body for reference to jira
+on:
+  pull_request:
+    branches:
+      - release_**
+jobs:
+  pr-check:
+    if: github.repository_owner == 'ansible' && github.repository != 'awx'
+    name: Scan PR description for JIRA links
+    runs-on: ubuntu-latest
+    permissions:
+      packages: write
+      contents: read
+    steps:
+      - name: Check for JIRA lines
+        env:
+          PR_BODY: ${{ github.event.pull_request.body }}
+        run: |
+          echo "$PR_BODY" | grep "JIRA: None" > no_jira
+          echo "$PR_BODY" | grep "JIRA: https://.*[0-9]+"> jira
+          exit 0
+        # We exit 0 and set the shell to prevent the returns from the greps from failing this step
+        # See https://docs.github.com/en/actions/using-workflows/workflow-syntax-for-github-actions#exit-codes-and-error-action-preference
+        shell: bash {0}
+
+      - name: Check for exactly one item
+        run: |
+          if [ $(cat no_jira jira | wc -l) != 1 ] ; then
+            echo "The PR body must contain exactly one of [ 'JIRA: None' or 'JIRA: <one or more links>' ]"
+            echo "We counted $(cat no_jira jira | wc -l)"
+            exit 255;
+          else
+            exit 0;
+          fi

DATA_MIGRATION.md

@@ -4,6 +4,6 @@
 
 Early versions of AWX did not support seamless upgrades between major versions and required the use of a backup and restore tool to perform upgrades.
 
-Users who wish to upgrade modern AWX installations should follow the instructions at:
+As of version 18.0, `awx-operator` is the preferred install/upgrade method. Users who wish to upgrade modern AWX installations should follow the instructions at:
 
-https://github.com/ansible/awx/blob/devel/INSTALL.md#upgrading-from-previous-versions
+https://github.com/ansible/awx-operator/blob/devel/docs/upgrade/upgrading.md

ISSUES.md

@@ -31,7 +31,7 @@ If your issue isn't considered high priority, then please be patient as it may t
 
 `state:needs_info` The issue needs more information. This could be more debug output, more specifics out the system such as version information. Any detail that is currently preventing this issue from moving forward. This should be considered a blocked state.
 
-`state:needs_review` The issue/pull request needs to be reviewed by other maintainers and contributors. This is usually used when there is a question out to another maintainer or when a person is less familar with an area of the code base the issue is for.
+`state:needs_review` The issue/pull request needs to be reviewed by other maintainers and contributors. This is usually used when there is a question out to another maintainer or when a person is less familiar with an area of the code base the issue is for.
 
 `state:needs_revision` More commonly used on pull requests, this state represents that there are changes that are being waited on.
 

Makefile

@@ -1,6 +1,7 @@
 -include awx/ui_next/Makefile
 
 PYTHON := $(notdir $(shell for i in python3.9 python3; do command -v $$i; done|sed 1q))
+SHELL := bash
 DOCKER_COMPOSE ?= docker-compose
 OFFICIAL ?= no
 NODE ?= node
@@ -8,7 +9,7 @@ NPM_BIN ?= npm
 CHROMIUM_BIN=/tmp/chrome-linux/chrome
 GIT_BRANCH ?= $(shell git rev-parse --abbrev-ref HEAD)
 MANAGEMENT_COMMAND ?= awx-manage
-VERSION := $(shell $(PYTHON) tools/scripts/scm_version.py)
+VERSION ?= $(shell $(PYTHON) tools/scripts/scm_version.py)
 
 # ansible-test requires semver compatable version, so we allow overrides to hack it
 COLLECTION_VERSION ?= $(shell $(PYTHON) tools/scripts/scm_version.py | cut -d . -f 1-3)
@@ -27,6 +28,8 @@ COLLECTION_TEMPLATE_VERSION ?= false
 # NOTE: This defaults the container image version to the branch that's active
 COMPOSE_TAG ?= $(GIT_BRANCH)
 MAIN_NODE_TYPE ?= hybrid
+# If set to true docker-compose will also start a pgbouncer instance and use it
+PGBOUNCER ?= false
 # If set to true docker-compose will also start a keycloak instance
 KEYCLOAK ?= false
 # If set to true docker-compose will also start an ldap instance
@@ -37,6 +40,8 @@ SPLUNK ?= false
 PROMETHEUS ?= false
 # If set to true docker-compose will also start a grafana instance
 GRAFANA ?= false
+# If set to true docker-compose will also start a hashicorp vault instance
+VAULT ?= false
 # If set to true docker-compose will also start a tacacs+ instance
 TACACS ?= false
 
@@ -52,7 +57,7 @@ RECEPTOR_IMAGE ?= quay.io/ansible/receptor:devel
 
 # Python packages to install only from source (not from binary wheels)
 # Comma separated list
-SRC_ONLY_PKGS ?= cffi,pycparser,psycopg2,twilio
+SRC_ONLY_PKGS ?= cffi,pycparser,psycopg,twilio
 # These should be upgraded in the AWX and Ansible venv before attempting
 # to install the actual requirements
 VENV_BOOTSTRAP ?= pip==21.2.4 setuptools==65.6.3 setuptools_scm[toml]==7.0.5 wheel==0.38.4
@@ -267,11 +272,11 @@ run-wsrelay:
 	$(PYTHON) manage.py run_wsrelay
 
 ## Start the heartbeat process in background in development environment.
-run-heartbeet:
+run-ws-heartbeat:
 	@if [ "$(VENV_BASE)" ]; then \
 		. $(VENV_BASE)/awx/bin/activate; \
 	fi; \
-	$(PYTHON) manage.py run_heartbeet
+	$(PYTHON) manage.py run_ws_heartbeat
 
 reports:
 	mkdir -p $@
@@ -520,15 +525,20 @@ docker-compose-sources: .git/hooks/pre-commit
 	    -e control_plane_node_count=$(CONTROL_PLANE_NODE_COUNT) \
 	    -e execution_node_count=$(EXECUTION_NODE_COUNT) \
 	    -e minikube_container_group=$(MINIKUBE_CONTAINER_GROUP) \
+	    -e enable_pgbouncer=$(PGBOUNCER) \
 	    -e enable_keycloak=$(KEYCLOAK) \
 	    -e enable_ldap=$(LDAP) \
 	    -e enable_splunk=$(SPLUNK) \
 	    -e enable_prometheus=$(PROMETHEUS) \
 	    -e enable_grafana=$(GRAFANA) \
+	    -e enable_vault=$(VAULT) \
 	    -e enable_tacacs=$(TACACS) \
             $(EXTRA_SOURCES_ANSIBLE_OPTS)
 
 docker-compose: awx/projects docker-compose-sources
+	ansible-galaxy install --ignore-certs -r tools/docker-compose/ansible/requirements.yml;
+	ansible-playbook -i tools/docker-compose/inventory tools/docker-compose/ansible/initialize_containers.yml \
+	  -e enable_vault=$(VAULT);
 	$(DOCKER_COMPOSE) -f tools/docker-compose/_sources/docker-compose.yml $(COMPOSE_OPTS) up $(COMPOSE_UP_OPTS) --remove-orphans
 
 docker-compose-credential-plugins: awx/projects docker-compose-sources
@@ -580,7 +590,7 @@ docker-clean:
 	-$(foreach image_id,$(shell docker images --filter=reference='*/*/*awx_devel*' --filter=reference='*/*awx_devel*' --filter=reference='*awx_devel*' -aq),docker rmi --force $(image_id);)
 
 docker-clean-volumes: docker-compose-clean docker-compose-container-group-clean
-	docker volume rm -f tools_awx_db tools_grafana_storage tools_prometheus_storage $(docker volume ls --filter name=tools_redis_socket_ -q)
+	docker volume rm -f tools_awx_db tools_vault_1 tools_grafana_storage tools_prometheus_storage $(docker volume ls --filter name=tools_redis_socket_ -q)
 
 docker-refresh: docker-clean docker-compose
 

awx/api/generics.py

@@ -232,7 +232,8 @@ class APIView(views.APIView):
 
         response = super(APIView, self).finalize_response(request, response, *args, **kwargs)
         time_started = getattr(self, 'time_started', None)
-        response['X-API-Product-Version'] = get_awx_version()
+        if request.user.is_authenticated:
+            response['X-API-Product-Version'] = get_awx_version()
         response['X-API-Product-Name'] = server_product_name()
 
         response['X-API-Node'] = settings.CLUSTER_HOST_ID

awx/api/serializers.py

@@ -1629,8 +1629,8 @@ class ProjectUpdateDetailSerializer(ProjectUpdateSerializer):
         fields = ('*', 'host_status_counts', 'playbook_counts')
 
     def get_playbook_counts(self, obj):
-        task_count = obj.project_update_events.filter(event='playbook_on_task_start').count()
-        play_count = obj.project_update_events.filter(event='playbook_on_play_start').count()
+        task_count = obj.get_event_queryset().filter(event='playbook_on_task_start').count()
+        play_count = obj.get_event_queryset().filter(event='playbook_on_play_start').count()
 
         data = {'play_count': play_count, 'task_count': task_count}
 
@@ -4686,12 +4686,11 @@ class BulkJobNodeSerializer(WorkflowJobNodeSerializer):
     # many-to-many fields
     credentials = serializers.ListField(child=serializers.IntegerField(min_value=1), required=False)
     labels = serializers.ListField(child=serializers.IntegerField(min_value=1), required=False)
-    # TODO: Use instance group role added via PR 13584(once merged), for now everything related to instance group is commented
-    # instance_groups = serializers.ListField(child=serializers.IntegerField(min_value=1), required=False)
+    instance_groups = serializers.ListField(child=serializers.IntegerField(min_value=1), required=False)
 
     class Meta:
         model = WorkflowJobNode
-        fields = ('*', 'credentials', 'labels')  # m2m fields are not canonical for WJ nodes, TODO: add instance_groups once supported
+        fields = ('*', 'credentials', 'labels', 'instance_groups')  # m2m fields are not canonical for WJ nodes
 
     def validate(self, attrs):
         return super(LaunchConfigurationBaseSerializer, self).validate(attrs)
@@ -4751,21 +4750,21 @@ class BulkJobLaunchSerializer(serializers.Serializer):
         requested_use_execution_environments = {job['execution_environment'] for job in attrs['jobs'] if 'execution_environment' in job}
         requested_use_credentials = set()
         requested_use_labels = set()
-        # requested_use_instance_groups = set()
+        requested_use_instance_groups = set()
         for job in attrs['jobs']:
             for cred in job.get('credentials', []):
                 requested_use_credentials.add(cred)
             for label in job.get('labels', []):
                 requested_use_labels.add(label)
-            # for instance_group in job.get('instance_groups', []):
-            #     requested_use_instance_groups.add(instance_group)
+            for instance_group in job.get('instance_groups', []):
+                requested_use_instance_groups.add(instance_group)
 
         key_to_obj_map = {
             "unified_job_template": {obj.id: obj for obj in UnifiedJobTemplate.objects.filter(id__in=requested_ujts)},
             "inventory": {obj.id: obj for obj in Inventory.objects.filter(id__in=requested_use_inventories)},
             "credentials": {obj.id: obj for obj in Credential.objects.filter(id__in=requested_use_credentials)},
             "labels": {obj.id: obj for obj in Label.objects.filter(id__in=requested_use_labels)},
-            # "instance_groups": {obj.id: obj for obj in InstanceGroup.objects.filter(id__in=requested_use_instance_groups)},
+            "instance_groups": {obj.id: obj for obj in InstanceGroup.objects.filter(id__in=requested_use_instance_groups)},
             "execution_environment": {obj.id: obj for obj in ExecutionEnvironment.objects.filter(id__in=requested_use_execution_environments)},
         }
 
@@ -4792,7 +4791,7 @@ class BulkJobLaunchSerializer(serializers.Serializer):
 
         self.check_list_permission(Credential, requested_use_credentials, 'use_role')
         self.check_list_permission(Label, requested_use_labels)
-        # self.check_list_permission(InstanceGroup, requested_use_instance_groups)  # TODO: change to use_role for conflict
+        self.check_list_permission(InstanceGroup, requested_use_instance_groups)  # TODO: change to use_role for conflict
         self.check_list_permission(ExecutionEnvironment, requested_use_execution_environments)  # TODO: change if roles introduced
 
         jobs_object = self.get_objectified_jobs(attrs, key_to_obj_map)
@@ -4839,7 +4838,7 @@ class BulkJobLaunchSerializer(serializers.Serializer):
         node_m2m_object_types_to_through_model = {
             'credentials': WorkflowJobNode.credentials.through,
             'labels': WorkflowJobNode.labels.through,
-            # 'instance_groups': WorkflowJobNode.instance_groups.through,
+            'instance_groups': WorkflowJobNode.instance_groups.through,
         }
         node_deferred_attr_names = (
             'limit',
@@ -4892,9 +4891,9 @@ class BulkJobLaunchSerializer(serializers.Serializer):
                 if field_name in node_m2m_objects[node_identifier] and field_name == 'labels':
                     for label in node_m2m_objects[node_identifier][field_name]:
                         through_model_objects.append(through_model(label=label, workflowjobnode=node_m2m_objects[node_identifier]['node']))
-                # if obj_type in node_m2m_objects[node_identifier] and obj_type == 'instance_groups':
-                #     for instance_group in node_m2m_objects[node_identifier][obj_type]:
-                #         through_model_objects.append(through_model(instancegroup=instance_group, workflowjobnode=node_m2m_objects[node_identifier]['node']))
+                if field_name in node_m2m_objects[node_identifier] and field_name == 'instance_groups':
+                    for instance_group in node_m2m_objects[node_identifier][field_name]:
+                        through_model_objects.append(through_model(instancegroup=instance_group, workflowjobnode=node_m2m_objects[node_identifier]['node']))
             if through_model_objects:
                 through_model.objects.bulk_create(through_model_objects)
 
@@ -5436,7 +5435,7 @@ class InstanceSerializer(BaseSerializer):
         res = super(InstanceSerializer, self).get_related(obj)
         res['jobs'] = self.reverse('api:instance_unified_jobs_list', kwargs={'pk': obj.pk})
         res['instance_groups'] = self.reverse('api:instance_instance_groups_list', kwargs={'pk': obj.pk})
-        if settings.IS_K8S and obj.node_type in (Instance.Types.EXECUTION,):
+        if obj.node_type in [Instance.Types.EXECUTION, Instance.Types.HOP]:
             res['install_bundle'] = self.reverse('api:instance_install_bundle', kwargs={'pk': obj.pk})
         res['peers'] = self.reverse('api:instance_peers_list', kwargs={"pk": obj.pk})
         if self.context['request'].user.is_superuser or self.context['request'].user.is_system_auditor:

awx/api/swagger.py

@@ -1,16 +1,10 @@
-import json
 import warnings
 
-from coreapi.document import Object, Link
-
-from rest_framework import exceptions
 from rest_framework.permissions import AllowAny
-from rest_framework.renderers import CoreJSONRenderer
-from rest_framework.response import Response
 from rest_framework.schemas import SchemaGenerator, AutoSchema as DRFAuthSchema
-from rest_framework.views import APIView
 
-from rest_framework_swagger import renderers
+from drf_yasg.views import get_schema_view
+from drf_yasg import openapi
 
 
 class SuperUserSchemaGenerator(SchemaGenerator):
@@ -55,43 +49,15 @@ class AutoSchema(DRFAuthSchema):
         return description
 
 
-class SwaggerSchemaView(APIView):
-    _ignore_model_permissions = True
-    exclude_from_schema = True
-    permission_classes = [AllowAny]
-    renderer_classes = [CoreJSONRenderer, renderers.OpenAPIRenderer, renderers.SwaggerUIRenderer]
-
-    def get(self, request):
-        generator = SuperUserSchemaGenerator(title='Ansible Automation Platform controller API', patterns=None, urlconf=None)
-        schema = generator.get_schema(request=request)
-        # python core-api doesn't support the deprecation yet, so track it
-        # ourselves and return it in a response header
-        _deprecated = []
-
-        # By default, DRF OpenAPI serialization places all endpoints in
-        # a single node based on their root path (/api).  Instead, we want to
-        # group them by topic/tag so that they're categorized in the rendered
-        # output
-        document = schema._data.pop('api')
-        for path, node in document.items():
-            if isinstance(node, Object):
-                for action in node.values():
-                    topic = getattr(action, 'topic', None)
-                    if topic:
-                        schema._data.setdefault(topic, Object())
-                        schema._data[topic]._data[path] = node
-
-                    if isinstance(action, Object):
-                        for link in action.links.values():
-                            if link.deprecated:
-                                _deprecated.append(link.url)
-            elif isinstance(node, Link):
-                topic = getattr(node, 'topic', None)
-                if topic:
-                    schema._data.setdefault(topic, Object())
-                    schema._data[topic]._data[path] = node
-
-        if not schema:
-            raise exceptions.ValidationError('The schema generator did not return a schema Document')
-
-        return Response(schema, headers={'X-Deprecated-Paths': json.dumps(_deprecated)})
+schema_view = get_schema_view(
+    openapi.Info(
+        title="Snippets API",
+        default_version='v1',
+        description="Test description",
+        terms_of_service="https://www.google.com/policies/terms/",
+        contact=openapi.Contact(email="contact@snippets.local"),
+        license=openapi.License(name="BSD License"),
+    ),
+    public=True,
+    permission_classes=[AllowAny],
+)

awx/api/templates/instance_install_bundle/group_vars/all.yml

@@ -9,10 +9,10 @@ receptor_work_commands:
     params: worker
     allowruntimeparams: true
     verifysignature: true
-custom_worksign_public_keyfile: receptor/work-public-key.pem
+custom_worksign_public_keyfile: receptor/work_public_key.pem
 custom_tls_certfile: receptor/tls/receptor.crt
 custom_tls_keyfile: receptor/tls/receptor.key
-custom_ca_certfile: receptor/tls/ca/receptor-ca.crt
+custom_ca_certfile: receptor/tls/ca/mesh-CA.crt
 receptor_protocol: 'tcp'
 receptor_listener: true
 receptor_port: {{ instance.listener_port }}

awx/api/urls/urls.py

@@ -30,7 +30,7 @@ from awx.api.views import (
     OAuth2TokenList,
     ApplicationOAuth2TokenList,
     OAuth2ApplicationDetail,
-    # HostMetricSummaryMonthlyList, # It will be enabled in future version of the AWX
+    HostMetricSummaryMonthlyList,
 )
 
 from awx.api.views.bulk import (
@@ -123,8 +123,7 @@ v2_urls = [
     re_path(r'^constructed_inventories/', include(constructed_inventory_urls)),
     re_path(r'^hosts/', include(host_urls)),
     re_path(r'^host_metrics/', include(host_metric_urls)),
-    # It will be enabled in future version of the AWX
-    # re_path(r'^host_metric_summary_monthly/$', HostMetricSummaryMonthlyList.as_view(), name='host_metric_summary_monthly_list'),
+    re_path(r'^host_metric_summary_monthly/$', HostMetricSummaryMonthlyList.as_view(), name='host_metric_summary_monthly_list'),
     re_path(r'^groups/', include(group_urls)),
     re_path(r'^inventory_sources/', include(inventory_source_urls)),
     re_path(r'^inventory_updates/', include(inventory_update_urls)),
@@ -167,10 +166,13 @@ urlpatterns = [
 ]
 if MODE == 'development':
     # Only include these if we are in the development environment
-    from awx.api.swagger import SwaggerSchemaView
-
-    urlpatterns += [re_path(r'^swagger/$', SwaggerSchemaView.as_view(), name='swagger_view')]
+    from awx.api.swagger import schema_view
 
     from awx.api.urls.debug import urls as debug_urls
 
     urlpatterns += [re_path(r'^debug/', include(debug_urls))]
+    urlpatterns += [
+        re_path(r'^swagger(?P<format>\.json|\.yaml)/$', schema_view.without_ui(cache_timeout=0), name='schema-json'),
+        re_path(r'^swagger/$', schema_view.with_ui('swagger', cache_timeout=0), name='schema-swagger-ui'),
+        re_path(r'^redoc/$', schema_view.with_ui('redoc', cache_timeout=0), name='schema-redoc'),
+    ]

awx/api/views/__init__.py

@@ -1564,16 +1564,15 @@ class HostMetricDetail(RetrieveDestroyAPIView):
         return Response(status=status.HTTP_204_NO_CONTENT)
 
 
-# It will be enabled in future version of the AWX
-# class HostMetricSummaryMonthlyList(ListAPIView):
-#     name = _("Host Metrics Summary Monthly")
-#     model = models.HostMetricSummaryMonthly
-#     serializer_class = serializers.HostMetricSummaryMonthlySerializer
-#     permission_classes = (IsSystemAdminOrAuditor,)
-#     search_fields = ('date',)
-#
-#     def get_queryset(self):
-#         return self.model.objects.all()
+class HostMetricSummaryMonthlyList(ListAPIView):
+    name = _("Host Metrics Summary Monthly")
+    model = models.HostMetricSummaryMonthly
+    serializer_class = serializers.HostMetricSummaryMonthlySerializer
+    permission_classes = (IsSystemAdminOrAuditor,)
+    search_fields = ('date',)
+
+    def get_queryset(self):
+        return self.model.objects.all()
 
 
 class HostList(HostRelatedSearchMixin, ListCreateAPIView):

awx/api/views/bulk.py

@@ -1,5 +1,7 @@
 from collections import OrderedDict
 
+from django.utils.translation import gettext_lazy as _
+
 from rest_framework.permissions import IsAuthenticated
 from rest_framework.renderers import JSONRenderer
 from rest_framework.reverse import reverse
@@ -18,6 +20,9 @@ from awx.api import (
 
 
 class BulkView(APIView):
+    name = _('Bulk')
+    swagger_topic = 'Bulk'
+
     permission_classes = [IsAuthenticated]
     renderer_classes = [
         renderers.BrowsableAPIRenderer,

awx/api/views/instance_install_bundle.py

@@ -57,13 +57,11 @@ class InstanceInstallBundle(GenericAPIView):
 
         with io.BytesIO() as f:
             with tarfile.open(fileobj=f, mode='w:gz') as tar:
-                # copy /etc/receptor/tls/ca/receptor-ca.crt to receptor/tls/ca in the tar file
-                tar.add(
-                    os.path.realpath('/etc/receptor/tls/ca/receptor-ca.crt'), arcname=f"{instance_obj.hostname}_install_bundle/receptor/tls/ca/receptor-ca.crt"
-                )
+                # copy /etc/receptor/tls/ca/mesh-CA.crt to receptor/tls/ca in the tar file
+                tar.add(os.path.realpath('/etc/receptor/tls/ca/mesh-CA.crt'), arcname=f"{instance_obj.hostname}_install_bundle/receptor/tls/ca/mesh-CA.crt")
 
-                # copy /etc/receptor/signing/work-public-key.pem to receptor/work-public-key.pem
-                tar.add('/etc/receptor/signing/work-public-key.pem', arcname=f"{instance_obj.hostname}_install_bundle/receptor/work-public-key.pem")
+                # copy /etc/receptor/work_public_key.pem to receptor/work_public_key.pem
+                tar.add('/etc/receptor/work_public_key.pem', arcname=f"{instance_obj.hostname}_install_bundle/receptor/work_public_key.pem")
 
                 # generate and write the receptor key to receptor/tls/receptor.key in the tar file
                 key, cert = generate_receptor_tls(instance_obj)
@@ -161,14 +159,14 @@ def generate_receptor_tls(instance_obj):
         .sign(key, hashes.SHA256())
     )
 
-    # sign csr with the receptor ca key from /etc/receptor/ca/receptor-ca.key
-    with open('/etc/receptor/tls/ca/receptor-ca.key', 'rb') as f:
+    # sign csr with the receptor ca key from /etc/receptor/ca/mesh-CA.key
+    with open('/etc/receptor/tls/ca/mesh-CA.key', 'rb') as f:
         ca_key = serialization.load_pem_private_key(
             f.read(),
             password=None,
         )
 
-    with open('/etc/receptor/tls/ca/receptor-ca.crt', 'rb') as f:
+    with open('/etc/receptor/tls/ca/mesh-CA.crt', 'rb') as f:
         ca_cert = x509.load_pem_x509_certificate(f.read())
 
     cert = (

awx/api/views/root.py

@@ -20,6 +20,7 @@ from rest_framework import status
 
 import requests
 
+from awx import MODE
 from awx.api.generics import APIView
 from awx.conf.registry import settings_registry
 from awx.main.analytics import all_collectors
@@ -54,6 +55,8 @@ class ApiRootView(APIView):
         data['custom_logo'] = settings.CUSTOM_LOGO
         data['custom_login_info'] = settings.CUSTOM_LOGIN_INFO
         data['login_redirect_override'] = settings.LOGIN_REDIRECT_OVERRIDE
+        if MODE == 'development':
+            data['swagger'] = drf_reverse('api:schema-swagger-ui')
         return Response(data)
 
 
@@ -104,8 +107,7 @@ class ApiVersionRootView(APIView):
         data['groups'] = reverse('api:group_list', request=request)
         data['hosts'] = reverse('api:host_list', request=request)
         data['host_metrics'] = reverse('api:host_metric_list', request=request)
-        # It will be enabled in future version of the AWX
-        # data['host_metric_summary_monthly'] = reverse('api:host_metric_summary_monthly_list', request=request)
+        data['host_metric_summary_monthly'] = reverse('api:host_metric_summary_monthly_list', request=request)
         data['job_templates'] = reverse('api:job_template_list', request=request)
         data['jobs'] = reverse('api:job_list', request=request)
         data['ad_hoc_commands'] = reverse('api:ad_hoc_command_list', request=request)

awx/conf/apps.py

@@ -14,7 +14,7 @@ class ConfConfig(AppConfig):
     def ready(self):
         self.module.autodiscover()
 
-        if not set(sys.argv) & {'migrate', 'check_migrations'}:
+        if not set(sys.argv) & {'migrate', 'check_migrations', 'showmigrations'}:
             from .settings import SettingsWrapper
 
             SettingsWrapper.initialize()

awx/conf/migrations/0010_change_to_JSONField.py

@@ -0,0 +1,17 @@
+# Generated by Django 4.2 on 2023-06-09 19:51
+
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ('conf', '0009_rename_proot_settings'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='setting',
+            name='value',
+            field=models.JSONField(null=True),
+        ),
+    ]

awx/conf/models.py

@@ -8,7 +8,6 @@ import json
 from django.db import models
 
 # AWX
-from awx.main.fields import JSONBlob
 from awx.main.models.base import CreatedModifiedModel, prevent_search
 from awx.main.utils import encrypt_field
 from awx.conf import settings_registry
@@ -18,7 +17,7 @@ __all__ = ['Setting']
 
 class Setting(CreatedModifiedModel):
     key = models.CharField(max_length=255)
-    value = JSONBlob(null=True)
+    value = models.JSONField(null=True)
     user = prevent_search(models.ForeignKey('auth.User', related_name='settings', default=None, null=True, editable=False, on_delete=models.CASCADE))
 
     def __str__(self):

awx/main/access.py

@@ -366,9 +366,9 @@ class BaseAccess(object):
             report_violation = lambda message: None
         else:
             report_violation = lambda message: logger.warning(message)
-        if validation_info.get('trial', False) is True or validation_info['instance_count'] == 10:  # basic 10 license
+        if validation_info.get('trial', False) is True:
 
-            def report_violation(message):
+            def report_violation(message):  # noqa
                 raise PermissionDenied(message)
 
         if check_expiration and validation_info.get('time_remaining', None) is None:

awx/main/analytics/collectors.py

@@ -399,7 +399,10 @@ def _copy_table(table, query, path):
     file_path = os.path.join(path, table + '_table.csv')
     file = FileSplitter(filespec=file_path)
     with connection.cursor() as cursor:
-        cursor.copy_expert(query, file)
+        with cursor.copy(query) as copy:
+            while data := copy.read():
+                byte_data = bytes(data)
+                file.write(byte_data.decode())
     return file.file_list()
 
 
@@ -610,3 +613,20 @@ def host_metric_table(since, full_path, until, **kwargs):
         since.isoformat(), until.isoformat(), since.isoformat(), until.isoformat()
     )
     return _copy_table(table='host_metric', query=host_metric_query, path=full_path)
+
+
+@register('host_metric_summary_monthly_table', '1.0', format='csv', description=_('HostMetricSummaryMonthly export, full sync'), expensive=trivial_slicing)
+def host_metric_summary_monthly_table(since, full_path, **kwargs):
+    query = '''
+    COPY (SELECT main_hostmetricsummarymonthly.id,
+                 main_hostmetricsummarymonthly.date,
+                 main_hostmetricsummarymonthly.license_capacity,
+                 main_hostmetricsummarymonthly.license_consumed,
+                 main_hostmetricsummarymonthly.hosts_added,
+                 main_hostmetricsummarymonthly.hosts_deleted,
+                 main_hostmetricsummarymonthly.indirectly_managed_hosts
+          FROM main_hostmetricsummarymonthly
+          ORDER BY main_hostmetricsummarymonthly.id ASC) TO STDOUT WITH CSV HEADER
+    '''
+
+    return _copy_table(table='host_metric_summary_monthly', query=query, path=full_path)

awx/main/analytics/subsystem_metrics.py

@@ -209,6 +209,11 @@ class Metrics:
             SetFloatM('workflow_manager_recorded_timestamp', 'Unix timestamp when metrics were last recorded'),
             SetFloatM('workflow_manager_spawn_workflow_graph_jobs_seconds', 'Time spent spawning workflow tasks'),
             SetFloatM('workflow_manager_get_tasks_seconds', 'Time spent loading workflow tasks from db'),
+            # dispatcher subsystem metrics
+            SetIntM('dispatcher_pool_scale_up_events', 'Number of times local dispatcher scaled up a worker since startup'),
+            SetIntM('dispatcher_pool_active_task_count', 'Number of active tasks in the worker pool when last task was submitted'),
+            SetIntM('dispatcher_pool_max_worker_count', 'Highest number of workers in worker pool in last collection interval, about 20s'),
+            SetFloatM('dispatcher_availability', 'Fraction of time (in last collection interval) dispatcher was able to receive messages'),
         ]
         # turn metric list into dictionary with the metric name as a key
         self.METRICS = {}

awx/main/cache.py

@@ -0,0 +1,87 @@
+import functools
+
+from django.conf import settings
+from django.core.cache.backends.base import DEFAULT_TIMEOUT
+from django.core.cache.backends.redis import RedisCache
+
+from redis.exceptions import ConnectionError, ResponseError, TimeoutError
+import socket
+
+# This list comes from what django-redis ignores and the behavior we are trying
+# to retain while dropping the dependency on django-redis.
+IGNORED_EXCEPTIONS = (TimeoutError, ResponseError, ConnectionError, socket.timeout)
+
+CONNECTION_INTERRUPTED_SENTINEL = object()
+
+
+def optionally_ignore_exceptions(func=None, return_value=None):
+    if func is None:
+        return functools.partial(optionally_ignore_exceptions, return_value=return_value)
+
+    @functools.wraps(func)
+    def wrapper(*args, **kwargs):
+        try:
+            return func(*args, **kwargs)
+        except IGNORED_EXCEPTIONS as e:
+            if settings.DJANGO_REDIS_IGNORE_EXCEPTIONS:
+                return return_value
+            raise e.__cause__ or e
+
+    return wrapper
+
+
+class AWXRedisCache(RedisCache):
+    """
+    We just want to wrap the upstream RedisCache class so that we can ignore
+    the exceptions that it raises when the cache is unavailable.
+    """
+
+    @optionally_ignore_exceptions
+    def add(self, key, value, timeout=DEFAULT_TIMEOUT, version=None):
+        return super().add(key, value, timeout, version)
+
+    @optionally_ignore_exceptions(return_value=CONNECTION_INTERRUPTED_SENTINEL)
+    def _get(self, key, default=None, version=None):
+        return super().get(key, default, version)
+
+    def get(self, key, default=None, version=None):
+        value = self._get(key, default, version)
+        if value is CONNECTION_INTERRUPTED_SENTINEL:
+            return default
+        return value
+
+    @optionally_ignore_exceptions
+    def set(self, key, value, timeout=DEFAULT_TIMEOUT, version=None):
+        return super().set(key, value, timeout, version)
+
+    @optionally_ignore_exceptions
+    def touch(self, key, timeout=DEFAULT_TIMEOUT, version=None):
+        return super().touch(key, timeout, version)
+
+    @optionally_ignore_exceptions
+    def delete(self, key, version=None):
+        return super().delete(key, version)
+
+    @optionally_ignore_exceptions
+    def get_many(self, keys, version=None):
+        return super().get_many(keys, version)
+
+    @optionally_ignore_exceptions
+    def has_key(self, key, version=None):
+        return super().has_key(key, version)
+
+    @optionally_ignore_exceptions
+    def incr(self, key, delta=1, version=None):
+        return super().incr(key, delta, version)
+
+    @optionally_ignore_exceptions
+    def set_many(self, data, timeout=DEFAULT_TIMEOUT, version=None):
+        return super().set_many(data, timeout, version)
+
+    @optionally_ignore_exceptions
+    def delete_many(self, keys, version=None):
+        return super().delete_many(keys, version)
+
+    @optionally_ignore_exceptions
+    def clear(self):
+        return super().clear()

awx/main/conf.py

@@ -94,6 +94,20 @@ register(
     category_slug='system',
 )
 
+register(
+    'CSRF_TRUSTED_ORIGINS',
+    default=[],
+    field_class=fields.StringListField,
+    label=_('CSRF Trusted Origins List'),
+    help_text=_(
+        "If the service is behind a reverse proxy/load balancer, use this setting "
+        "to configure the schema://addresses from which the service should trust "
+        "Origin header values. "
+    ),
+    category=_('System'),
+    category_slug='system',
+)
+
 register(
     'LICENSE',
     field_class=fields.DictField,
@@ -684,11 +698,28 @@ register(
     field_class=fields.IntegerField,
     default=1,
     min_value=1,
-    label=_('Maximum disk persistance for external log aggregation (in GB)'),
+    label=_('Maximum disk persistence for external log aggregation (in GB)'),
     help_text=_(
         'Amount of data to store (in gigabytes) during an outage of '
         'the external log aggregator (defaults to 1). '
-        'Equivalent to the rsyslogd queue.maxdiskspace setting.'
+        'Equivalent to the rsyslogd queue.maxdiskspace setting for main_queue. '
+        'Notably, this is used for the rsyslogd main queue (for input messages).'
+    ),
+    category=_('Logging'),
+    category_slug='logging',
+)
+register(
+    'LOG_AGGREGATOR_ACTION_MAX_DISK_USAGE_GB',
+    field_class=fields.IntegerField,
+    default=1,
+    min_value=1,
+    label=_('Maximum disk persistence for rsyslogd action queuing (in GB)'),
+    help_text=_(
+        'Amount of data to store (in gigabytes) if an rsyslog action takes time '
+        'to process an incoming message (defaults to 1). '
+        'Equivalent to the rsyslogd queue.maxdiskspace setting on the action (e.g. omhttp). '
+        'Like LOG_AGGREGATOR_MAX_DISK_USAGE_GB, it stores files in the directory specified '
+        'by LOG_AGGREGATOR_MAX_DISK_USAGE_PATH.'
     ),
     category=_('Logging'),
     category_slug='logging',
@@ -831,6 +862,55 @@ register(
     category_slug='system',
 )
 
+register(
+    'HOST_METRIC_SUMMARY_TASK_LAST_TS',
+    field_class=fields.DateTimeField,
+    label=_('Last computing date of HostMetricSummaryMonthly'),
+    allow_null=True,
+    category=_('System'),
+    category_slug='system',
+)
+
+register(
+    'AWX_CLEANUP_PATHS',
+    field_class=fields.BooleanField,
+    label=_('Enable or Disable tmp dir cleanup'),
+    default=True,
+    help_text=_('Enable or Disable TMP Dir cleanup'),
+    category=('Debug'),
+    category_slug='debug',
+)
+
+register(
+    'AWX_REQUEST_PROFILE',
+    field_class=fields.BooleanField,
+    label=_('Debug Web Requests'),
+    default=False,
+    help_text=_('Debug web request python timing'),
+    category=('Debug'),
+    category_slug='debug',
+)
+
+register(
+    'DEFAULT_CONTAINER_RUN_OPTIONS',
+    field_class=fields.StringListField,
+    label=_('Container Run Options'),
+    default=['--network', 'slirp4netns:enable_ipv6=true'],
+    help_text=_("List of options to pass to podman run example: ['--network', 'slirp4netns:enable_ipv6=true', '--log-level', 'debug']"),
+    category=('Jobs'),
+    category_slug='jobs',
+)
+
+register(
+    'RECEPTOR_RELEASE_WORK',
+    field_class=fields.BooleanField,
+    label=_('Release Receptor Work'),
+    default=True,
+    help_text=_('Release receptor work'),
+    category=('Debug'),
+    category_slug='debug',
+)
+
 
 def logging_validate(serializer, attrs):
     if not serializer.instance or not hasattr(serializer.instance, 'LOG_AGGREGATOR_HOST') or not hasattr(serializer.instance, 'LOG_AGGREGATOR_TYPE'):

awx/main/credential_plugins/aws_secretsmanager.py

@@ -0,0 +1,65 @@
+import boto3
+from botocore.exceptions import ClientError
+
+from .plugin import CredentialPlugin
+from django.utils.translation import gettext_lazy as _
+
+
+secrets_manager_inputs = {
+    'fields': [
+        {
+            'id': 'aws_access_key',
+            'label': _('AWS Access Key'),
+            'type': 'string',
+        },
+        {
+            'id': 'aws_secret_key',
+            'label': _('AWS Secret Key'),
+            'type': 'string',
+            'secret': True,
+        },
+    ],
+    'metadata': [
+        {
+            'id': 'region_name',
+            'label': _('AWS Secrets Manager Region'),
+            'type': 'string',
+            'help_text': _('Region which the secrets manager is located'),
+        },
+        {
+            'id': 'secret_name',
+            'label': _('AWS Secret Name'),
+            'type': 'string',
+        },
+    ],
+    'required': ['aws_access_key', 'aws_secret_key', 'region_name', 'secret_name'],
+}
+
+
+def aws_secretsmanager_backend(**kwargs):
+    secret_name = kwargs['secret_name']
+    region_name = kwargs['region_name']
+    aws_secret_access_key = kwargs['aws_secret_key']
+    aws_access_key_id = kwargs['aws_access_key']
+
+    session = boto3.session.Session()
+    client = session.client(
+        service_name='secretsmanager', region_name=region_name, aws_secret_access_key=aws_secret_access_key, aws_access_key_id=aws_access_key_id
+    )
+
+    try:
+        get_secret_value_response = client.get_secret_value(SecretId=secret_name)
+    except ClientError as e:
+        raise e
+    # Secrets Manager decrypts the secret value using the associated KMS CMK
+    # Depending on whether the secret was a string or binary, only one of these fields will be populated
+    if 'SecretString' in get_secret_value_response:
+        secret = get_secret_value_response['SecretString']
+
+    else:
+        secret = get_secret_value_response['SecretBinary']
+
+    return secret
+
+
+aws_secretmanager_plugin = CredentialPlugin('AWS Secrets Manager lookup', inputs=secrets_manager_inputs, backend=aws_secretsmanager_backend)

awx/main/credential_plugins/hashivault.py

@@ -265,6 +265,8 @@ def kv_backend(**kwargs):
 
     if secret_key:
         try:
+            if (secret_key != 'data') and (secret_key not in json['data']) and ('data' in json['data']):
+                return json['data']['data'][secret_key]
             return json['data'][secret_key]
         except KeyError:
             raise RuntimeError('{} is not present at {}'.format(secret_key, secret_path))

awx/main/credential_plugins/tss.py

@@ -50,7 +50,7 @@ tss_inputs = {
 
 
 def tss_backend(**kwargs):
-    if 'domain' in kwargs:
+    if kwargs.get("domain"):
         authorizer = DomainPasswordGrantAuthorizer(kwargs['server_url'], kwargs['username'], kwargs['password'], kwargs['domain'])
     else:
         authorizer = PasswordGrantAuthorizer(kwargs['server_url'], kwargs['username'], kwargs['password'])

awx/main/dispatch/__init__.py

@@ -1,5 +1,5 @@
 import os
-import psycopg2
+import psycopg
 import select
 
 from contextlib import contextmanager
@@ -40,8 +40,12 @@ def get_task_queuename():
 
 
 class PubSub(object):
-    def __init__(self, conn):
+    def __init__(self, conn, select_timeout=None):
         self.conn = conn
+        if select_timeout is None:
+            self.select_timeout = 5
+        else:
+            self.select_timeout = select_timeout
 
     def listen(self, channel):
         with self.conn.cursor() as cur:
@@ -55,25 +59,42 @@ class PubSub(object):
         with self.conn.cursor() as cur:
             cur.execute('SELECT pg_notify(%s, %s);', (channel, payload))
 
-    def events(self, select_timeout=5, yield_timeouts=False):
+    @staticmethod
+    def current_notifies(conn):
+        """
+        Altered version of .notifies method from psycopg library
+        This removes the outer while True loop so that we only process
+        queued notifications
+        """
+        with conn.lock:
+            try:
+                ns = conn.wait(psycopg.generators.notifies(conn.pgconn))
+            except psycopg.errors._NO_TRACEBACK as ex:
+                raise ex.with_traceback(None)
+        enc = psycopg._encodings.pgconn_encoding(conn.pgconn)
+        for pgn in ns:
+            n = psycopg.connection.Notify(pgn.relname.decode(enc), pgn.extra.decode(enc), pgn.be_pid)
+            yield n
+
+    def events(self, yield_timeouts=False):
         if not self.conn.autocommit:
             raise RuntimeError('Listening for events can only be done in autocommit mode')
 
         while True:
-            if select.select([self.conn], [], [], select_timeout) == NOT_READY:
+            if select.select([self.conn], [], [], self.select_timeout) == NOT_READY:
                 if yield_timeouts:
                     yield None
             else:
-                self.conn.poll()
-                while self.conn.notifies:
-                    yield self.conn.notifies.pop(0)
+                notification_generator = self.current_notifies(self.conn)
+                for notification in notification_generator:
+                    yield notification
 
     def close(self):
         self.conn.close()
 
 
 @contextmanager
-def pg_bus_conn(new_connection=False):
+def pg_bus_conn(new_connection=False, select_timeout=None):
     '''
     Any listeners probably want to establish a new database connection,
     separate from the Django connection used for queries, because that will prevent
@@ -89,9 +110,8 @@ def pg_bus_conn(new_connection=False):
         conf['OPTIONS'] = conf.get('OPTIONS', {}).copy()
         # Modify the application name to distinguish from other connections the process might use
         conf['OPTIONS']['application_name'] = get_application_name(settings.CLUSTER_HOST_ID, function='listener')
-        conn = psycopg2.connect(dbname=conf['NAME'], host=conf['HOST'], user=conf['USER'], password=conf['PASSWORD'], port=conf['PORT'], **conf['OPTIONS'])
-        # Django connection.cursor().connection doesn't have autocommit=True on by default
-        conn.set_session(autocommit=True)
+        connection_data = f"dbname={conf['NAME']} host={conf['HOST']} user={conf['USER']} password={conf['PASSWORD']} port={conf['PORT']}"
+        conn = psycopg.connect(connection_data, autocommit=True, **conf['OPTIONS'])
     else:
         if pg_connection.connection is None:
             pg_connection.connect()
@@ -99,7 +119,7 @@ def pg_bus_conn(new_connection=False):
             raise RuntimeError('Unexpectedly could not connect to postgres for pg_notify actions')
         conn = pg_connection.connection
 
-    pubsub = PubSub(conn)
+    pubsub = PubSub(conn, select_timeout=select_timeout)
     yield pubsub
     if new_connection:
         conn.close()

awx/main/dispatch/control.py

@@ -40,6 +40,9 @@ class Control(object):
     def cancel(self, task_ids, *args, **kwargs):
         return self.control_with_reply('cancel', *args, extra_data={'task_ids': task_ids}, **kwargs)
 
+    def schedule(self, *args, **kwargs):
+        return self.control_with_reply('schedule', *args, **kwargs)
+
     @classmethod
     def generate_reply_queue_name(cls):
         return f"reply_to_{str(uuid.uuid4()).replace('-','_')}"
@@ -52,14 +55,14 @@ class Control(object):
         if not connection.get_autocommit():
             raise RuntimeError('Control-with-reply messages can only be done in autocommit mode')
 
-        with pg_bus_conn() as conn:
+        with pg_bus_conn(select_timeout=timeout) as conn:
             conn.listen(reply_queue)
             send_data = {'control': command, 'reply_to': reply_queue}
             if extra_data:
                 send_data.update(extra_data)
             conn.notify(self.queuename, json.dumps(send_data))
 
-            for reply in conn.events(select_timeout=timeout, yield_timeouts=True):
+            for reply in conn.events(yield_timeouts=True):
                 if reply is None:
                     logger.error(f'{self.service} did not reply within {timeout}s')
                     raise RuntimeError(f"{self.service} did not reply within {timeout}s")

awx/main/dispatch/periodic.py

@@ -1,57 +1,142 @@
 import logging
-import os
 import time
-from multiprocessing import Process
+import yaml
+from datetime import datetime
 
-from django.conf import settings
-from django.db import connections
-from schedule import Scheduler
-from django_guid import set_guid
-from django_guid.utils import generate_guid
-
-from awx.main.dispatch.worker import TaskWorker
-from awx.main.utils.db import set_connection_name
 
 logger = logging.getLogger('awx.main.dispatch.periodic')
 
 
-class Scheduler(Scheduler):
-    def run_continuously(self):
-        idle_seconds = max(1, min(self.jobs).period.total_seconds() / 2)
+class ScheduledTask:
+    """
+    Class representing schedules, very loosely modeled after python schedule library Job
+    the idea of this class is to:
+     - only deal in relative times (time since the scheduler global start)
+     - only deal in integer math for target runtimes, but float for current relative time
 
-        def run():
-            ppid = os.getppid()
-            logger.warning('periodic beat started')
+    Missed schedule policy:
+    Invariant target times are maintained, meaning that if interval=10s offset=0
+    and it runs at t=7s, then it calls for next run in 3s.
+    However, if a complete interval has passed, that is counted as a missed run,
+    and missed runs are abandoned (no catch-up runs).
+    """
 
-            set_connection_name('periodic')  # set application_name to distinguish from other dispatcher processes
+    def __init__(self, name: str, data: dict):
+        # parameters need for schedule computation
+        self.interval = int(data['schedule'].total_seconds())
+        self.offset = 0  # offset relative to start time this schedule begins
+        self.index = 0  # number of periods of the schedule that has passed
 
-            while True:
-                if os.getppid() != ppid:
-                    # if the parent PID changes, this process has been orphaned
-                    # via e.g., segfault or sigkill, we should exit too
-                    pid = os.getpid()
-                    logger.warning(f'periodic beat exiting gracefully pid:{pid}')
-                    raise SystemExit()
-                try:
-                    for conn in connections.all():
-                        # If the database connection has a hiccup, re-establish a new
-                        # connection
-                        conn.close_if_unusable_or_obsolete()
-                    set_guid(generate_guid())
-                    self.run_pending()
-                except Exception:
-                    logger.exception('encountered an error while scheduling periodic tasks')
-                time.sleep(idle_seconds)
+        # parameters that do not affect scheduling logic
+        self.last_run = None  # time of last run, only used for debug
+        self.completed_runs = 0  # number of times schedule is known to run
+        self.name = name
+        self.data = data  # used by caller to know what to run
 
-        process = Process(target=run)
-        process.daemon = True
-        process.start()
+    @property
+    def next_run(self):
+        "Time until the next run with t=0 being the global_start of the scheduler class"
+        return (self.index + 1) * self.interval + self.offset
+
+    def due_to_run(self, relative_time):
+        return bool(self.next_run <= relative_time)
+
+    def expected_runs(self, relative_time):
+        return int((relative_time - self.offset) / self.interval)
+
+    def mark_run(self, relative_time):
+        self.last_run = relative_time
+        self.completed_runs += 1
+        new_index = self.expected_runs(relative_time)
+        if new_index > self.index + 1:
+            logger.warning(f'Missed {new_index - self.index - 1} schedules of {self.name}')
+        self.index = new_index
+
+    def missed_runs(self, relative_time):
+        "Number of times job was supposed to ran but failed to, only used for debug"
+        missed_ct = self.expected_runs(relative_time) - self.completed_runs
+        # if this is currently due to run do not count that as a missed run
+        if missed_ct and self.due_to_run(relative_time):
+            missed_ct -= 1
+        return missed_ct
 
 
-def run_continuously():
-    scheduler = Scheduler()
-    for task in settings.CELERYBEAT_SCHEDULE.values():
-        apply_async = TaskWorker.resolve_callable(task['task']).apply_async
-        total_seconds = task['schedule'].total_seconds()
-        scheduler.every(total_seconds).seconds.do(apply_async)
-    scheduler.run_continuously()
+class Scheduler:
+    def __init__(self, schedule):
+        """
+        Expects schedule in the form of a dictionary like
+        {
+            'job1': {'schedule': timedelta(seconds=50), 'other': 'stuff'}
+        }
+        Only the schedule nearest-second value is used for scheduling,
+        the rest of the data is for use by the caller to know what to run.
+        """
+        self.jobs = [ScheduledTask(name, data) for name, data in schedule.items()]
+        min_interval = min(job.interval for job in self.jobs)
+        num_jobs = len(self.jobs)
+
+        # this is intentionally oppioniated against spammy schedules
+        # a core goal is to spread out the scheduled tasks (for worker management)
+        # and high-frequency schedules just do not work with that
+        if num_jobs > min_interval:
+            raise RuntimeError(f'Number of schedules ({num_jobs}) is more than the shortest schedule interval ({min_interval} seconds).')
+
+        # even space out jobs over the base interval
+        for i, job in enumerate(self.jobs):
+            job.offset = (i * min_interval) // num_jobs
+
+        # internally times are all referenced relative to startup time, add grace period
+        self.global_start = time.time() + 2.0
+
+    def get_and_mark_pending(self):
+        relative_time = time.time() - self.global_start
+        to_run = []
+        for job in self.jobs:
+            if job.due_to_run(relative_time):
+                to_run.append(job)
+                logger.debug(f'scheduler found {job.name} to run, {relative_time - job.next_run} seconds after target')
+                job.mark_run(relative_time)
+        return to_run
+
+    def time_until_next_run(self):
+        relative_time = time.time() - self.global_start
+        next_job = min(self.jobs, key=lambda j: j.next_run)
+        delta = next_job.next_run - relative_time
+        if delta <= 0.1:
+            # careful not to give 0 or negative values to the select timeout, which has unclear interpretation
+            logger.warning(f'Scheduler next run of {next_job.name} is {-delta} seconds in the past')
+            return 0.1
+        elif delta > 20.0:
+            logger.warning(f'Scheduler next run unexpectedly over 20 seconds in future: {delta}')
+            return 20.0
+        logger.debug(f'Scheduler next run is {next_job.name} in {delta} seconds')
+        return delta
+
+    def debug(self, *args, **kwargs):
+        data = dict()
+        data['title'] = 'Scheduler status'
+
+        now = datetime.fromtimestamp(time.time()).strftime('%Y-%m-%d %H:%M:%S UTC')
+        start_time = datetime.fromtimestamp(self.global_start).strftime('%Y-%m-%d %H:%M:%S UTC')
+        relative_time = time.time() - self.global_start
+        data['started_time'] = start_time
+        data['current_time'] = now
+        data['current_time_relative'] = round(relative_time, 3)
+        data['total_schedules'] = len(self.jobs)
+
+        data['schedule_list'] = dict(
+            [
+                (
+                    job.name,
+                    dict(
+                        last_run_seconds_ago=round(relative_time - job.last_run, 3) if job.last_run else None,
+                        next_run_in_seconds=round(job.next_run - relative_time, 3),
+                        offset_in_seconds=job.offset,
+                        completed_runs=job.completed_runs,
+                        missed_runs=job.missed_runs(relative_time),
+                    ),
+                )
+                for job in sorted(self.jobs, key=lambda job: job.interval)
+            ]
+        )
+        return yaml.safe_dump(data, default_flow_style=False, sort_keys=False)

awx/main/dispatch/pool.py

@@ -339,6 +339,17 @@ class AutoscalePool(WorkerPool):
         # but if the task takes longer than the time defined here, we will force it to stop here
         self.task_manager_timeout = settings.TASK_MANAGER_TIMEOUT + settings.TASK_MANAGER_TIMEOUT_GRACE_PERIOD
 
+        # initialize some things for subsystem metrics periodic gathering
+        # the AutoscalePool class does not save these to redis directly, but reports via produce_subsystem_metrics
+        self.scale_up_ct = 0
+        self.worker_count_max = 0
+
+    def produce_subsystem_metrics(self, metrics_object):
+        metrics_object.set('dispatcher_pool_scale_up_events', self.scale_up_ct)
+        metrics_object.set('dispatcher_pool_active_task_count', sum(len(w.managed_tasks) for w in self.workers))
+        metrics_object.set('dispatcher_pool_max_worker_count', self.worker_count_max)
+        self.worker_count_max = len(self.workers)
+
     @property
     def should_grow(self):
         if len(self.workers) < self.min_workers:
@@ -406,16 +417,16 @@ class AutoscalePool(WorkerPool):
                 # the task manager to never do more work
                 current_task = w.current_task
                 if current_task and isinstance(current_task, dict):
-                    endings = ['tasks.task_manager', 'tasks.dependency_manager', 'tasks.workflow_manager']
+                    endings = ('tasks.task_manager', 'tasks.dependency_manager', 'tasks.workflow_manager')
                     current_task_name = current_task.get('task', '')
-                    if any(current_task_name.endswith(e) for e in endings):
+                    if current_task_name.endswith(endings):
                         if 'started' not in current_task:
                             w.managed_tasks[current_task['uuid']]['started'] = time.time()
                         age = time.time() - current_task['started']
                         w.managed_tasks[current_task['uuid']]['age'] = age
                         if age > self.task_manager_timeout:
-                            logger.error(f'{current_task_name} has held the advisory lock for {age}, sending SIGTERM to {w.pid}')
-                            os.kill(w.pid, signal.SIGTERM)
+                            logger.error(f'{current_task_name} has held the advisory lock for {age}, sending SIGUSR1 to {w.pid}')
+                            os.kill(w.pid, signal.SIGUSR1)
 
         for m in orphaned:
             # if all the workers are dead, spawn at least one
@@ -443,7 +454,12 @@ class AutoscalePool(WorkerPool):
             idx = random.choice(range(len(self.workers)))
             return idx, self.workers[idx]
         else:
-            return super(AutoscalePool, self).up()
+            self.scale_up_ct += 1
+            ret = super(AutoscalePool, self).up()
+            new_worker_ct = len(self.workers)
+            if new_worker_ct > self.worker_count_max:
+                self.worker_count_max = new_worker_ct
+            return ret
 
     def write(self, preferred_queue, body):
         if 'guid' in body:

awx/main/dispatch/publish.py

@@ -73,15 +73,15 @@ class task:
                 return cls.apply_async(args, kwargs)
 
             @classmethod
-            def apply_async(cls, args=None, kwargs=None, queue=None, uuid=None, **kw):
+            def get_async_body(cls, args=None, kwargs=None, uuid=None, **kw):
+                """
+                Get the python dict to become JSON data in the pg_notify message
+                This same message gets passed over the dispatcher IPC queue to workers
+                If a task is submitted to a multiprocessing pool, skipping pg_notify, this might be used directly
+                """
                 task_id = uuid or str(uuid4())
                 args = args or []
                 kwargs = kwargs or {}
-                queue = queue or getattr(cls.queue, 'im_func', cls.queue)
-                if not queue:
-                    msg = f'{cls.name}: Queue value required and may not be None'
-                    logger.error(msg)
-                    raise ValueError(msg)
                 obj = {'uuid': task_id, 'args': args, 'kwargs': kwargs, 'task': cls.name, 'time_pub': time.time()}
                 guid = get_guid()
                 if guid:
@@ -89,6 +89,16 @@ class task:
                 if bind_kwargs:
                     obj['bind_kwargs'] = bind_kwargs
                 obj.update(**kw)
+                return obj
+
+            @classmethod
+            def apply_async(cls, args=None, kwargs=None, queue=None, uuid=None, **kw):
+                queue = queue or getattr(cls.queue, 'im_func', cls.queue)
+                if not queue:
+                    msg = f'{cls.name}: Queue value required and may not be None'
+                    logger.error(msg)
+                    raise ValueError(msg)
+                obj = cls.get_async_body(args=args, kwargs=kwargs, uuid=uuid, **kw)
                 if callable(queue):
                     queue = queue()
                 if not is_testing():
@@ -116,4 +126,5 @@ class task:
         setattr(fn, 'name', cls.name)
         setattr(fn, 'apply_async', cls.apply_async)
         setattr(fn, 'delay', cls.delay)
+        setattr(fn, 'get_async_body', cls.get_async_body)
         return fn

awx/main/dispatch/worker/base.py

@@ -7,18 +7,21 @@ import signal
 import sys
 import redis
 import json
-import psycopg2
+import psycopg
 import time
 from uuid import UUID
 from queue import Empty as QueueEmpty
+from datetime import timedelta
 
 from django import db
 from django.conf import settings
 
 from awx.main.dispatch.pool import WorkerPool
+from awx.main.dispatch.periodic import Scheduler
 from awx.main.dispatch import pg_bus_conn
 from awx.main.utils.common import log_excess_runtime
 from awx.main.utils.db import set_connection_name
+import awx.main.analytics.subsystem_metrics as s_metrics
 
 if 'run_callback_receiver' in sys.argv:
     logger = logging.getLogger('awx.main.commands.run_callback_receiver')
@@ -63,10 +66,12 @@ class AWXConsumerBase(object):
     def control(self, body):
         logger.warning(f'Received control signal:\n{body}')
         control = body.get('control')
-        if control in ('status', 'running', 'cancel'):
+        if control in ('status', 'schedule', 'running', 'cancel'):
             reply_queue = body['reply_to']
             if control == 'status':
                 msg = '\n'.join([self.listening_on, self.pool.debug()])
+            if control == 'schedule':
+                msg = self.scheduler.debug()
             elif control == 'running':
                 msg = []
                 for worker in self.pool.workers:
@@ -92,16 +97,11 @@ class AWXConsumerBase(object):
         else:
             logger.error('unrecognized control message: {}'.format(control))
 
-    def process_task(self, body):
+    def dispatch_task(self, body):
+        """This will place the given body into a worker queue to run method decorated as a task"""
         if isinstance(body, dict):
             body['time_ack'] = time.time()
 
-        if 'control' in body:
-            try:
-                return self.control(body)
-            except Exception:
-                logger.exception(f"Exception handling control message: {body}")
-                return
         if len(self.pool):
             if "uuid" in body and body['uuid']:
                 try:
@@ -115,15 +115,24 @@ class AWXConsumerBase(object):
         self.pool.write(queue, body)
         self.total_messages += 1
 
+    def process_task(self, body):
+        """Routes the task details in body as either a control task or a task-task"""
+        if 'control' in body:
+            try:
+                return self.control(body)
+            except Exception:
+                logger.exception(f"Exception handling control message: {body}")
+                return
+        self.dispatch_task(body)
+
     @log_excess_runtime(logger)
     def record_statistics(self):
         if time.time() - self.last_stats > 1:  # buffer stat recording to once per second
             try:
                 self.redis.set(f'awx_{self.name}_statistics', self.pool.debug())
-                self.last_stats = time.time()
             except Exception:
                 logger.exception(f"encountered an error communicating with redis to store {self.name} statistics")
-                self.last_stats = time.time()
+            self.last_stats = time.time()
 
     def run(self, *args, **kwargs):
         signal.signal(signal.SIGINT, self.stop)
@@ -142,29 +151,72 @@ class AWXConsumerRedis(AWXConsumerBase):
     def run(self, *args, **kwargs):
         super(AWXConsumerRedis, self).run(*args, **kwargs)
         self.worker.on_start()
+        logger.info(f'Callback receiver started with pid={os.getpid()}')
+        db.connection.close()  # logs use database, so close connection
 
         while True:
-            logger.debug(f'{os.getpid()} is alive')
             time.sleep(60)
 
 
 class AWXConsumerPG(AWXConsumerBase):
-    def __init__(self, *args, **kwargs):
+    def __init__(self, *args, schedule=None, **kwargs):
         super().__init__(*args, **kwargs)
         self.pg_max_wait = settings.DISPATCHER_DB_DOWNTOWN_TOLLERANCE
         # if no successful loops have ran since startup, then we should fail right away
         self.pg_is_down = True  # set so that we fail if we get database errors on startup
-        self.pg_down_time = time.time() - self.pg_max_wait  # allow no grace period
-        self.last_cleanup = time.time()
+        init_time = time.time()
+        self.pg_down_time = init_time - self.pg_max_wait  # allow no grace period
+        self.last_cleanup = init_time
+        self.subsystem_metrics = s_metrics.Metrics(auto_pipe_execute=False)
+        self.last_metrics_gather = init_time
+        self.listen_cumulative_time = 0.0
+        if schedule:
+            schedule = schedule.copy()
+        else:
+            schedule = {}
+        # add control tasks to be ran at regular schedules
+        # NOTE: if we run out of database connections, it is important to still run cleanup
+        # so that we scale down workers and free up connections
+        schedule['pool_cleanup'] = {'control': self.pool.cleanup, 'schedule': timedelta(seconds=60)}
+        # record subsystem metrics for the dispatcher
+        schedule['metrics_gather'] = {'control': self.record_metrics, 'schedule': timedelta(seconds=20)}
+        self.scheduler = Scheduler(schedule)
+
+    def record_metrics(self):
+        current_time = time.time()
+        self.pool.produce_subsystem_metrics(self.subsystem_metrics)
+        self.subsystem_metrics.set('dispatcher_availability', self.listen_cumulative_time / (current_time - self.last_metrics_gather))
+        self.subsystem_metrics.pipe_execute()
+        self.listen_cumulative_time = 0.0
+        self.last_metrics_gather = current_time
 
     def run_periodic_tasks(self):
-        self.record_statistics()  # maintains time buffer in method
+        """
+        Run general periodic logic, and return maximum time in seconds before
+        the next requested run
+        This may be called more often than that when events are consumed
+        so this should be very efficient in that
+        """
+        try:
+            self.record_statistics()  # maintains time buffer in method
+        except Exception as exc:
+            logger.warning(f'Failed to save dispatcher statistics {exc}')
 
-        if time.time() - self.last_cleanup > 60:  # same as cluster_node_heartbeat
-            # NOTE: if we run out of database connections, it is important to still run cleanup
-            # so that we scale down workers and free up connections
-            self.pool.cleanup()
-            self.last_cleanup = time.time()
+        for job in self.scheduler.get_and_mark_pending():
+            if 'control' in job.data:
+                try:
+                    job.data['control']()
+                except Exception:
+                    logger.exception(f'Error running control task {job.data}')
+            elif 'task' in job.data:
+                body = self.worker.resolve_callable(job.data['task']).get_async_body()
+                # bypasses pg_notify for scheduled tasks
+                self.dispatch_task(body)
+
+        self.pg_is_down = False
+        self.listen_start = time.time()
+
+        return self.scheduler.time_until_next_run()
 
     def run(self, *args, **kwargs):
         super(AWXConsumerPG, self).run(*args, **kwargs)
@@ -180,17 +232,21 @@ class AWXConsumerPG(AWXConsumerBase):
                     if init is False:
                         self.worker.on_start()
                         init = True
+                    # run_periodic_tasks run scheduled actions and gives time until next scheduled action
+                    # this is saved to the conn (PubSub) object in order to modify read timeout in-loop
+                    conn.select_timeout = self.run_periodic_tasks()
+                    # this is the main operational loop for awx-manage run_dispatcher
                     for e in conn.events(yield_timeouts=True):
+                        self.listen_cumulative_time += time.time() - self.listen_start  # for metrics
                         if e is not None:
                             self.process_task(json.loads(e.payload))
-                        self.run_periodic_tasks()
-                        self.pg_is_down = False
+                        conn.select_timeout = self.run_periodic_tasks()
                     if self.should_stop:
                         return
-            except psycopg2.InterfaceError:
+            except psycopg.InterfaceError:
                 logger.warning("Stale Postgres message bus connection, reconnecting")
                 continue
-            except (db.DatabaseError, psycopg2.OperationalError):
+            except (db.DatabaseError, psycopg.OperationalError):
                 # If we have attained stady state operation, tolerate short-term database hickups
                 if not self.pg_is_down:
                     logger.exception(f"Error consuming new events from postgres, will retry for {self.pg_max_wait} s")
@@ -232,8 +288,8 @@ class BaseWorker(object):
                     break
             except QueueEmpty:
                 continue
-            except Exception as e:
-                logger.error("Exception on worker {}, restarting: ".format(idx) + str(e))
+            except Exception:
+                logger.exception("Exception on worker {}, reconnecting: ".format(idx))
                 continue
             try:
                 for conn in db.connections.all():

awx/main/dispatch/worker/callback.py

@@ -191,7 +191,9 @@ class CallbackBrokerWorker(BaseWorker):
                             e._retry_count = retry_count
 
                             # special sanitization logic for postgres treatment of NUL 0x00 char
-                            if (retry_count == 1) and isinstance(exc_indv, ValueError) and ("\x00" in e.stdout):
+                            # This used to check the class of the exception but on the postgres3 upgrade it could appear
+                            #   as either DataError or ValueError, so now lets just try if its there.
+                            if (retry_count == 1) and ("\x00" in e.stdout):
                                 e.stdout = e.stdout.replace("\x00", "")
 
                             if retry_count >= self.INDIVIDUAL_EVENT_RETRIES:

awx/main/fields.py

@@ -67,10 +67,60 @@ def __enum_validate__(validator, enums, instance, schema):
 Draft4Validator.VALIDATORS['enum'] = __enum_validate__
 
 
+import logging
+
+logger = logging.getLogger('awx.main.fields')
+
+
 class JSONBlob(JSONField):
+    # Cringe... a JSONField that is back ended with a TextField.
+    # This field was a legacy custom field type that tl;dr; was a TextField
+    # Over the years, with Django upgrades, we were able to go to a JSONField instead of the custom field
+    # However, we didn't want to have large customers with millions of events to update from text to json during an upgrade
+    # So we keep this field type as backended with TextField.
     def get_internal_type(self):
         return "TextField"
 
+    # postgres uses a Jsonb field as the default backend
+    # with psycopg2 it was using a psycopg2._json.Json class internally
+    # with psycopg3 it uses a psycopg.types.json.Jsonb class internally
+    # The binary class was not compatible with a text field, so we are going to override these next two methods and ensure we are using a string
+
+    def from_db_value(self, value, expression, connection):
+        if value is None:
+            return value
+
+        if isinstance(value, str):
+            try:
+                return json.loads(value)
+            except Exception as e:
+                logger.error(f"Failed to load JSONField {self.name}: {e}")
+
+        return value
+
+    def get_db_prep_value(self, value, connection, prepared=False):
+        if not prepared:
+            value = self.get_prep_value(value)
+        try:
+            # Null characters are not allowed in text fields and JSONBlobs are JSON data but saved as text
+            # So we want to make sure we strip out any null characters also note, these "should" be escaped by the dumps process:
+            #     >>> my_obj = { 'test': '\x00' }
+            #     >>> import json
+            #     >>> json.dumps(my_obj)
+            #     '{"test": "\\u0000"}'
+            # But just to be safe, lets remove them if they are there. \x00 and \u0000 are the same:
+            #     >>> string = "\x00"
+            #     >>> "\u0000" in string
+            #     True
+            dumped_value = json.dumps(value)
+            if "\x00" in dumped_value:
+                dumped_value = dumped_value.replace("\x00", '')
+            return dumped_value
+        except Exception as e:
+            logger.error(f"Failed to dump JSONField {self.name}: {e} value: {value}")
+
+        return value
+
 
 # Based on AutoOneToOneField from django-annoying:
 # https://bitbucket.org/offline/django-annoying/src/a0de8b294db3/annoying/fields.py

awx/main/management/commands/cleanup_host_metrics.py

@@ -17,6 +17,6 @@ class Command(BaseCommand):
         months_ago = options.get('months-ago') or None
 
         if not months_ago:
-            months_ago = getattr(settings, 'CLEANUP_HOST_METRICS_THRESHOLD', 12)
+            months_ago = getattr(settings, 'CLEANUP_HOST_METRICS_SOFT_THRESHOLD', 12)
 
         HostMetric.cleanup_task(months_ago)

awx/main/management/commands/cleanup_jobs.py

@@ -17,10 +17,7 @@ from django.utils.timezone import now
 
 # AWX
 from awx.main.models import Job, AdHocCommand, ProjectUpdate, InventoryUpdate, SystemJob, WorkflowJob, Notification
-
-
-def unified_job_class_to_event_table_name(job_class):
-    return f'main_{job_class().event_class.__name__.lower()}'
+from awx.main.utils import unified_job_class_to_event_table_name
 
 
 def partition_table_name(job_class, dt):

awx/main/management/commands/host_metric_summary_monthly.py

@@ -0,0 +1,9 @@
+from django.core.management.base import BaseCommand
+from awx.main.tasks.host_metrics import HostMetricSummaryMonthlyTask
+
+
+class Command(BaseCommand):
+    help = 'Computing of HostMetricSummaryMonthly'
+
+    def handle(self, *args, **options):
+        HostMetricSummaryMonthlyTask().execute()

awx/main/management/commands/precreate_partitions.py

@@ -0,0 +1,27 @@
+from django.utils.timezone import now
+from django.core.management.base import BaseCommand, CommandParser
+from datetime import timedelta
+from awx.main.utils.common import create_partition, unified_job_class_to_event_table_name
+from awx.main.models import Job, SystemJob, ProjectUpdate, InventoryUpdate, AdHocCommand
+
+
+class Command(BaseCommand):
+    """Command used to precreate database partitions to avoid pg_dump locks"""
+
+    def add_arguments(self, parser: CommandParser) -> None:
+        parser.add_argument('--count', dest='count', action='store', help='The amount of hours of partitions to create', type=int, default=1)
+
+    def _create_partitioned_tables(self, count):
+        tables = list()
+        for model in (Job, SystemJob, ProjectUpdate, InventoryUpdate, AdHocCommand):
+            tables.append(unified_job_class_to_event_table_name(model))
+        start = now()
+        while count > 0:
+            for table in tables:
+                create_partition(table, start)
+                print(f'Created partitions for {table} {start}')
+            start = start + timedelta(hours=1)
+            count -= 1
+
+    def handle(self, **options):
+        self._create_partitioned_tables(count=options.get('count'))

awx/main/management/commands/provision_instance.py

@@ -35,7 +35,7 @@ class Command(BaseCommand):
 
             from awx.main.management.commands.register_queue import RegisterQueue
 
-            (changed, instance) = Instance.objects.register(ip_address=os.environ.get('MY_POD_IP'), node_type='control', uuid=settings.SYSTEM_UUID)
+            (changed, instance) = Instance.objects.register(ip_address=os.environ.get('MY_POD_IP'), node_type='control', node_uuid=settings.SYSTEM_UUID)
             RegisterQueue(settings.DEFAULT_CONTROL_PLANE_QUEUE_NAME, 100, 0, [], is_container_group=False).register()
             RegisterQueue(
                 settings.DEFAULT_EXECUTION_QUEUE_NAME,
@@ -48,7 +48,7 @@ class Command(BaseCommand):
                 max_concurrent_jobs=settings.DEFAULT_EXECUTION_QUEUE_MAX_CONCURRENT_JOBS,
             ).register()
         else:
-            (changed, instance) = Instance.objects.register(hostname=hostname, node_type=node_type, uuid=uuid)
+            (changed, instance) = Instance.objects.register(hostname=hostname, node_type=node_type, node_uuid=uuid)
         if changed:
             print("Successfully registered instance {}".format(hostname))
         else:

awx/main/management/commands/run_cache_clear.py

@@ -2,6 +2,7 @@ import logging
 import json
 
 from django.core.management.base import BaseCommand
+
 from awx.main.dispatch import pg_bus_conn
 from awx.main.dispatch.worker.task import TaskWorker
 
@@ -18,7 +19,7 @@ class Command(BaseCommand):
 
     def handle(self, *arg, **options):
         try:
-            with pg_bus_conn(new_connection=True) as conn:
+            with pg_bus_conn() as conn:
                 conn.listen("tower_settings_change")
                 for e in conn.events(yield_timeouts=True):
                     if e is not None:

awx/main/management/commands/run_dispatcher.py

@@ -4,28 +4,22 @@ import logging
 import yaml
 
 from django.conf import settings
-from django.core.cache import cache as django_cache
 from django.core.management.base import BaseCommand
-from django.db import connection as django_connection
 
 from awx.main.dispatch import get_task_queuename
 from awx.main.dispatch.control import Control
 from awx.main.dispatch.pool import AutoscalePool
 from awx.main.dispatch.worker import AWXConsumerPG, TaskWorker
-from awx.main.dispatch import periodic
 
 logger = logging.getLogger('awx.main.dispatch')
 
 
-def construct_bcast_queue_name(common_name):
-    return common_name + '_' + settings.CLUSTER_HOST_ID
-
-
 class Command(BaseCommand):
     help = 'Launch the task dispatcher'
 
     def add_arguments(self, parser):
         parser.add_argument('--status', dest='status', action='store_true', help='print the internal state of any running dispatchers')
+        parser.add_argument('--schedule', dest='schedule', action='store_true', help='print the current status of schedules being ran by dispatcher')
         parser.add_argument('--running', dest='running', action='store_true', help='print the UUIDs of any tasked managed by this dispatcher')
         parser.add_argument(
             '--reload',
@@ -47,6 +41,9 @@ class Command(BaseCommand):
         if options.get('status'):
             print(Control('dispatcher').status())
             return
+        if options.get('schedule'):
+            print(Control('dispatcher').schedule())
+            return
         if options.get('running'):
             print(Control('dispatcher').running())
             return
@@ -63,21 +60,11 @@ class Command(BaseCommand):
             print(Control('dispatcher').cancel(cancel_data))
             return
 
-        # It's important to close these because we're _about_ to fork, and we
-        # don't want the forked processes to inherit the open sockets
-        # for the DB and cache connections (that way lies race conditions)
-        django_connection.close()
-        django_cache.close()
-
-        # spawn a daemon thread to periodically enqueues scheduled tasks
-        # (like the node heartbeat)
-        periodic.run_continuously()
-
         consumer = None
 
         try:
             queues = ['tower_broadcast_all', 'tower_settings_change', get_task_queuename()]
-            consumer = AWXConsumerPG('dispatcher', TaskWorker(), queues, AutoscalePool(min_workers=4))
+            consumer = AWXConsumerPG('dispatcher', TaskWorker(), queues, AutoscalePool(min_workers=4), schedule=settings.CELERYBEAT_SCHEDULE)
             consumer.run()
         except KeyboardInterrupt:
             logger.debug('Terminating Task Dispatcher')

awx/main/management/commands/run_heartbeet.py

@@ -1,74 +0,0 @@
-import json
-import logging
-import os
-import time
-import signal
-import sys
-
-from django.core.management.base import BaseCommand
-from django.conf import settings
-
-from awx.main.dispatch import pg_bus_conn
-
-logger = logging.getLogger('awx.main.commands.run_heartbeet')
-
-
-class Command(BaseCommand):
-    help = 'Launch the web server beacon (heartbeet)'
-
-    def print_banner(self):
-        heartbeet = r"""
-   **********   **********
- ************* *************
-*****************************
- ***********HEART***********
-  *************************
-     *******************
-       *************** _._
-         *********** /`._ `'.      __
-           *******   \  .\| \   _'`  `)
-             ***  (``_)  \| ).'` /`- /
-              *   `\ `;\_ `\\//`-'` /
-                    \ `'.'.| /  __/`
-                     `'--v_|/`'`
-                         __||-._
-                      /'` `-``  `'\\
-                     /         .'` )
-                     \  BEET  '    )
-                      \.          /
-                        '.     /'`
-                           `) |
-                            //
-                            '(.
-                             `\`.
-                               ``"""
-        print(heartbeet)
-
-    def construct_payload(self, action='online'):
-        payload = {
-            'hostname': settings.CLUSTER_HOST_ID,
-            'ip': os.environ.get('MY_POD_IP'),
-            'action': action,
-        }
-        return json.dumps(payload)
-
-    def notify_listener_and_exit(self, *args):
-        with pg_bus_conn(new_connection=False) as conn:
-            conn.notify('web_heartbeet', self.construct_payload(action='offline'))
-        sys.exit(0)
-
-    def do_hearbeat_loop(self):
-        with pg_bus_conn(new_connection=True) as conn:
-            while True:
-                logger.debug('Sending heartbeat')
-                conn.notify('web_heartbeet', self.construct_payload())
-                time.sleep(settings.BROADCAST_WEBSOCKET_BEACON_FROM_WEB_RATE_SECONDS)
-
-    def handle(self, *arg, **options):
-        self.print_banner()
-        signal.signal(signal.SIGTERM, self.notify_listener_and_exit)
-        signal.signal(signal.SIGINT, self.notify_listener_and_exit)
-
-        # Note: We don't really try any reconnect logic to pg_notify here,
-        # just let supervisor restart if we fail.
-        self.do_hearbeat_loop()

awx/main/management/commands/run_rsyslog_configurer.py

@@ -22,7 +22,7 @@ class Command(BaseCommand):
 
     def handle(self, *arg, **options):
         try:
-            with pg_bus_conn(new_connection=True) as conn:
+            with pg_bus_conn() as conn:
                 conn.listen("rsyslog_configurer")
                 # reconfigure rsyslog on start up
                 reconfigure_rsyslog()

awx/main/management/commands/run_ws_heartbeat.py

@@ -0,0 +1,45 @@
+import json
+import logging
+import os
+import time
+import signal
+import sys
+
+from django.core.management.base import BaseCommand
+from django.conf import settings
+
+from awx.main.dispatch import pg_bus_conn
+
+logger = logging.getLogger('awx.main.commands.run_ws_heartbeat')
+
+
+class Command(BaseCommand):
+    help = 'Launch the web server beacon (ws_heartbeat)'
+
+    def construct_payload(self, action='online'):
+        payload = {
+            'hostname': settings.CLUSTER_HOST_ID,
+            'ip': os.environ.get('MY_POD_IP'),
+            'action': action,
+        }
+        return json.dumps(payload)
+
+    def notify_listener_and_exit(self, *args):
+        with pg_bus_conn(new_connection=False) as conn:
+            conn.notify('web_ws_heartbeat', self.construct_payload(action='offline'))
+        sys.exit(0)
+
+    def do_heartbeat_loop(self):
+        while True:
+            with pg_bus_conn() as conn:
+                logger.debug('Sending heartbeat')
+                conn.notify('web_ws_heartbeat', self.construct_payload())
+            time.sleep(settings.BROADCAST_WEBSOCKET_BEACON_FROM_WEB_RATE_SECONDS)
+
+    def handle(self, *arg, **options):
+        signal.signal(signal.SIGTERM, self.notify_listener_and_exit)
+        signal.signal(signal.SIGINT, self.notify_listener_and_exit)
+
+        # Note: We don't really try any reconnect logic to pg_notify here,
+        # just let supervisor restart if we fail.
+        self.do_heartbeat_loop()

awx/main/managers.py

@@ -2,6 +2,7 @@
 # All Rights Reserved.
 
 import logging
+import uuid
 from django.db import models
 from django.conf import settings
 from django.db.models.functions import Lower
@@ -114,7 +115,7 @@ class InstanceManager(models.Manager):
             return node[0]
         raise RuntimeError("No instance found with the current cluster host id")
 
-    def register(self, uuid=None, hostname=None, ip_address=None, node_type='hybrid', defaults=None):
+    def register(self, node_uuid=None, hostname=None, ip_address=None, node_type='hybrid', defaults=None):
         if not hostname:
             hostname = settings.CLUSTER_HOST_ID
 
@@ -131,8 +132,8 @@ class InstanceManager(models.Manager):
                         logger.warning("IP address {0} conflict detected, ip address unset for host {1}.".format(ip_address, other_hostname))
 
             # Return existing instance that matches hostname or UUID (default to UUID)
-            if uuid is not None and uuid != UUID_DEFAULT and self.filter(uuid=uuid).exists():
-                instance = self.filter(uuid=uuid)
+            if node_uuid is not None and node_uuid != UUID_DEFAULT and self.filter(uuid=node_uuid).exists():
+                instance = self.filter(uuid=node_uuid)
             else:
                 # if instance was not retrieved by uuid and hostname was, use the hostname
                 instance = self.filter(hostname=hostname)
@@ -170,9 +171,7 @@ class InstanceManager(models.Manager):
             }
             if defaults is not None:
                 create_defaults.update(defaults)
-            uuid_option = {}
-            if uuid is not None:
-                uuid_option = {'uuid': uuid}
+            uuid_option = {'uuid': node_uuid if node_uuid is not None else uuid.uuid4()}
             if node_type == 'execution' and 'version' not in create_defaults:
                 create_defaults['version'] = RECEPTOR_PENDING
             instance = self.create(hostname=hostname, ip_address=ip_address, node_type=node_type, **create_defaults, **uuid_option)

awx/main/migrations/0001_initial.py

@@ -9,13 +9,11 @@ from django.db import migrations, models
 import django.utils.timezone
 import django.db.models.deletion
 from django.conf import settings
-import taggit.managers
 import awx.main.fields
 
 
 class Migration(migrations.Migration):
     dependencies = [
-        ('taggit', '0002_auto_20150616_2121'),
         ('contenttypes', '0002_remove_content_type_name'),
         migrations.swappable_dependency(settings.AUTH_USER_MODEL),
     ]
@@ -184,12 +182,6 @@ class Migration(migrations.Migration):
                         null=True,
                     ),
                 ),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-                    ),
-                ),
             ],
             options={
                 'ordering': ('kind', 'name'),
@@ -529,12 +521,6 @@ class Migration(migrations.Migration):
                         null=True,
                     ),
                 ),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-                    ),
-                ),
                 ('users', models.ManyToManyField(related_name='organizations', to=settings.AUTH_USER_MODEL, blank=True)),
             ],
             options={
@@ -589,12 +575,6 @@ class Migration(migrations.Migration):
                         null=True,
                     ),
                 ),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-                    ),
-                ),
             ],
         ),
         migrations.CreateModel(
@@ -644,12 +624,6 @@ class Migration(migrations.Migration):
                         null=True,
                     ),
                 ),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-                    ),
-                ),
             ],
             options={
                 'ordering': ['-next_run'],
@@ -687,12 +661,6 @@ class Migration(migrations.Migration):
                     ),
                 ),
                 ('organization', models.ForeignKey(related_name='teams', on_delete=django.db.models.deletion.SET_NULL, to='main.Organization', null=True)),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-                    ),
-                ),
                 ('users', models.ManyToManyField(related_name='teams', to=settings.AUTH_USER_MODEL, blank=True)),
             ],
             options={
@@ -1267,13 +1235,6 @@ class Migration(migrations.Migration):
                 null=True,
             ),
         ),
-        migrations.AddField(
-            model_name='unifiedjobtemplate',
-            name='tags',
-            field=taggit.managers.TaggableManager(
-                to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-            ),
-        ),
         migrations.AddField(
             model_name='unifiedjob',
             name='created_by',
@@ -1319,13 +1280,6 @@ class Migration(migrations.Migration):
             name='schedule',
             field=models.ForeignKey(on_delete=django.db.models.deletion.SET_NULL, default=None, editable=False, to='main.Schedule', null=True),
         ),
-        migrations.AddField(
-            model_name='unifiedjob',
-            name='tags',
-            field=taggit.managers.TaggableManager(
-                to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-            ),
-        ),
         migrations.AddField(
             model_name='unifiedjob',
             name='unified_job_template',
@@ -1370,13 +1324,6 @@ class Migration(migrations.Migration):
                 help_text='Organization containing this inventory.',
             ),
         ),
-        migrations.AddField(
-            model_name='inventory',
-            name='tags',
-            field=taggit.managers.TaggableManager(
-                to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-            ),
-        ),
         migrations.AddField(
             model_name='host',
             name='inventory',
@@ -1407,13 +1354,6 @@ class Migration(migrations.Migration):
                 null=True,
             ),
         ),
-        migrations.AddField(
-            model_name='host',
-            name='tags',
-            field=taggit.managers.TaggableManager(
-                to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-            ),
-        ),
         migrations.AddField(
             model_name='group',
             name='hosts',
@@ -1441,13 +1381,6 @@ class Migration(migrations.Migration):
             name='parents',
             field=models.ManyToManyField(related_name='children', to='main.Group', blank=True),
         ),
-        migrations.AddField(
-            model_name='group',
-            name='tags',
-            field=taggit.managers.TaggableManager(
-                to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-            ),
-        ),
         migrations.AddField(
             model_name='custominventoryscript',
             name='organization',
@@ -1459,13 +1392,6 @@ class Migration(migrations.Migration):
                 null=True,
             ),
         ),
-        migrations.AddField(
-            model_name='custominventoryscript',
-            name='tags',
-            field=taggit.managers.TaggableManager(
-                to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-            ),
-        ),
         migrations.AddField(
             model_name='credential',
             name='team',

awx/main/migrations/0002_squashed_v300_release.py

@@ -12,8 +12,6 @@ import django.db.models.deletion
 from django.conf import settings
 from django.utils.timezone import now
 
-import taggit.managers
-
 
 def create_system_job_templates(apps, schema_editor):
     """
@@ -125,7 +123,6 @@ class Migration(migrations.Migration):
     ]
 
     dependencies = [
-        ('taggit', '0002_auto_20150616_2121'),
         ('contenttypes', '0002_remove_content_type_name'),
         migrations.swappable_dependency(settings.AUTH_USER_MODEL),
         ('main', '0001_initial'),
@@ -256,12 +253,6 @@ class Migration(migrations.Migration):
                     'organization',
                     models.ForeignKey(related_name='notification_templates', on_delete=django.db.models.deletion.SET_NULL, to='main.Organization', null=True),
                 ),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-                    ),
-                ),
             ],
         ),
         migrations.AddField(
@@ -721,12 +712,6 @@ class Migration(migrations.Migration):
                         help_text='Organization this label belongs to.',
                     ),
                 ),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-                    ),
-                ),
             ],
             options={
                 'ordering': ('organization', 'name'),

awx/main/migrations/0006_v320_release.py

@@ -2,13 +2,9 @@
 # Python
 from __future__ import unicode_literals
 
-# Psycopg2
-from psycopg2.extensions import AsIs
-
 # Django
 from django.db import connection, migrations, models, OperationalError, ProgrammingError
 from django.conf import settings
-import taggit.managers
 
 # AWX
 import awx.main.fields
@@ -136,8 +132,8 @@ class Migration(migrations.Migration):
             ),
         ),
         migrations.RunSQL(
-            [("CREATE INDEX host_ansible_facts_default_gin ON %s USING gin" "(ansible_facts jsonb_path_ops);", [AsIs(Host._meta.db_table)])],
-            [('DROP INDEX host_ansible_facts_default_gin;', None)],
+            sql="CREATE INDEX host_ansible_facts_default_gin ON {} USING gin(ansible_facts jsonb_path_ops);".format(Host._meta.db_table),
+            reverse_sql='DROP INDEX host_ansible_facts_default_gin;',
         ),
         # SCM file-based inventories
         migrations.AddField(
@@ -320,10 +316,6 @@ class Migration(migrations.Migration):
             model_name='permission',
             name='project',
         ),
-        migrations.RemoveField(
-            model_name='permission',
-            name='tags',
-        ),
         migrations.RemoveField(
             model_name='permission',
             name='team',
@@ -513,12 +505,6 @@ class Migration(migrations.Migration):
                         null=True,
                     ),
                 ),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        to='taggit.Tag', through='taggit.TaggedItem', blank=True, help_text='A comma-separated list of tags.', verbose_name='Tags'
-                    ),
-                ),
             ],
             options={
                 'ordering': ('kind', 'name'),

awx/main/migrations/0067_v350_credential_plugins.py

@@ -4,7 +4,6 @@ from __future__ import unicode_literals
 from django.conf import settings
 from django.db import migrations, models
 import django.db.models.deletion
-import taggit.managers
 
 # AWX
 import awx.main.fields
@@ -20,7 +19,6 @@ def setup_tower_managed_defaults(apps, schema_editor):
 class Migration(migrations.Migration):
     dependencies = [
         migrations.swappable_dependency(settings.AUTH_USER_MODEL),
-        ('taggit', '0002_auto_20150616_2121'),
         ('main', '0066_v350_inventorysource_custom_virtualenv'),
     ]
 
@@ -60,12 +58,6 @@ class Migration(migrations.Migration):
                     'source_credential',
                     models.ForeignKey(null=True, on_delete=django.db.models.deletion.CASCADE, related_name='target_input_sources', to='main.Credential'),
                 ),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        blank=True, help_text='A comma-separated list of tags.', through='taggit.TaggedItem', to='taggit.Tag', verbose_name='Tags'
-                    ),
-                ),
                 (
                     'target_credential',
                     models.ForeignKey(null=True, on_delete=django.db.models.deletion.CASCADE, related_name='input_sources', to='main.Credential'),

awx/main/migrations/0124_execution_environments.py

@@ -4,12 +4,10 @@ from django.conf import settings
 from django.db import migrations, models
 import django.db.models.deletion
 import django.db.models.expressions
-import taggit.managers
 
 
 class Migration(migrations.Migration):
     dependencies = [
-        ('taggit', '0003_taggeditem_add_unique_index'),
         migrations.swappable_dependency(settings.AUTH_USER_MODEL),
         ('main', '0123_drop_hg_support'),
     ]
@@ -69,12 +67,6 @@ class Migration(migrations.Migration):
                         to='main.Organization',
                     ),
                 ),
-                (
-                    'tags',
-                    taggit.managers.TaggableManager(
-                        blank=True, help_text='A comma-separated list of tags.', through='taggit.TaggedItem', to='taggit.Tag', verbose_name='Tags'
-                    ),
-                ),
             ],
             options={
                 'ordering': (django.db.models.expressions.OrderBy(django.db.models.expressions.F('organization_id'), nulls_first=True), 'image'),

awx/main/migrations/0144_event_partitions.py

@@ -30,7 +30,7 @@ def migrate_event_data(apps, schema_editor):
             # otherwise, the schema changes we would make on the old jobevents table
             # (namely, dropping the primary key constraint) would cause the migration
             # to suffer a serious performance degradation
-            cursor.execute(f'CREATE TABLE tmp_{tblname} ' f'(LIKE _unpartitioned_{tblname} INCLUDING ALL)')
+            cursor.execute(f'CREATE TABLE tmp_{tblname} (LIKE _unpartitioned_{tblname} INCLUDING ALL)')
 
             # drop primary key constraint; in a partioned table
             # constraints must include the partition key itself
@@ -48,7 +48,7 @@ def migrate_event_data(apps, schema_editor):
             cursor.execute(f'DROP TABLE tmp_{tblname}')
 
             # recreate primary key constraint
-            cursor.execute(f'ALTER TABLE ONLY {tblname} ' f'ADD CONSTRAINT {tblname}_pkey_new PRIMARY KEY (id, job_created);')
+            cursor.execute(f'ALTER TABLE ONLY {tblname} ADD CONSTRAINT {tblname}_pkey_new PRIMARY KEY (id, job_created);')
 
     with connection.cursor() as cursor:
         """

awx/main/migrations/0185_move_JSONBlob_to_JSONField.py

@@ -0,0 +1,277 @@
+# Generated by Django 4.2.3 on 2023-08-02 13:18
+
+import awx.main.models.notifications
+from django.db import migrations, models
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ('main', '0184_django_indexes'),
+        ('conf', '0010_change_to_JSONField'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='instancegroup',
+            name='policy_instance_list',
+            field=models.JSONField(
+                blank=True, default=list, help_text='List of exact-match Instances that will always be automatically assigned to this group'
+            ),
+        ),
+        migrations.AlterField(
+            model_name='jobtemplate',
+            name='survey_spec',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.AlterField(
+            model_name='notificationtemplate',
+            name='messages',
+            field=models.JSONField(
+                blank=True,
+                default=awx.main.models.notifications.NotificationTemplate.default_messages,
+                help_text='Optional custom messages for notification template.',
+                null=True,
+            ),
+        ),
+        migrations.AlterField(
+            model_name='notificationtemplate',
+            name='notification_configuration',
+            field=models.JSONField(default=dict),
+        ),
+        migrations.AlterField(
+            model_name='project',
+            name='inventory_files',
+            field=models.JSONField(
+                blank=True,
+                default=list,
+                editable=False,
+                help_text='Suggested list of content that could be Ansible inventory in the project',
+                verbose_name='Inventory Files',
+            ),
+        ),
+        migrations.AlterField(
+            model_name='project',
+            name='playbook_files',
+            field=models.JSONField(blank=True, default=list, editable=False, help_text='List of playbooks found in the project', verbose_name='Playbook Files'),
+        ),
+        migrations.AlterField(
+            model_name='schedule',
+            name='char_prompts',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.AlterField(
+            model_name='schedule',
+            name='survey_passwords',
+            field=models.JSONField(blank=True, default=dict, editable=False),
+        ),
+        migrations.AlterField(
+            model_name='workflowjobtemplate',
+            name='char_prompts',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.AlterField(
+            model_name='workflowjobtemplate',
+            name='survey_spec',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.AlterField(
+            model_name='workflowjobtemplatenode',
+            name='char_prompts',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.AlterField(
+            model_name='workflowjobtemplatenode',
+            name='survey_passwords',
+            field=models.JSONField(blank=True, default=dict, editable=False),
+        ),
+        # These are potentially a problem.  Move the existing fields
+        # aside while pretending like they've been deleted, then add
+        # in fresh empty fields.  Make the old fields nullable where
+        # needed while we are at it, so that new rows don't hit
+        # IntegrityError.  We'll do the data migration out-of-band
+        # using a task.
+        migrations.RunSQL(  # Already nullable
+            "ALTER TABLE main_activitystream RENAME deleted_actor TO deleted_actor_old;",
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='activitystream',
+                    name='deleted_actor',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='activitystream',
+            name='deleted_actor',
+            field=models.JSONField(null=True),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_activitystream RENAME setting TO setting_old;
+            ALTER TABLE main_activitystream ALTER COLUMN setting_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='activitystream',
+                    name='setting',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='activitystream',
+            name='setting',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_job RENAME survey_passwords TO survey_passwords_old;
+            ALTER TABLE main_job ALTER COLUMN survey_passwords_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='job',
+                    name='survey_passwords',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='job',
+            name='survey_passwords',
+            field=models.JSONField(blank=True, default=dict, editable=False),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_joblaunchconfig RENAME char_prompts TO char_prompts_old;
+            ALTER TABLE main_joblaunchconfig ALTER COLUMN char_prompts_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='joblaunchconfig',
+                    name='char_prompts',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='joblaunchconfig',
+            name='char_prompts',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_joblaunchconfig RENAME survey_passwords TO survey_passwords_old;
+            ALTER TABLE main_joblaunchconfig ALTER COLUMN survey_passwords_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='joblaunchconfig',
+                    name='survey_passwords',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='joblaunchconfig',
+            name='survey_passwords',
+            field=models.JSONField(blank=True, default=dict, editable=False),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_notification RENAME body TO body_old;
+            ALTER TABLE main_notification ALTER COLUMN body_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='notification',
+                    name='body',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='notification',
+            name='body',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_unifiedjob RENAME job_env TO job_env_old;
+            ALTER TABLE main_unifiedjob ALTER COLUMN job_env_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='unifiedjob',
+                    name='job_env',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='unifiedjob',
+            name='job_env',
+            field=models.JSONField(blank=True, default=dict, editable=False),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_workflowjob RENAME char_prompts TO char_prompts_old;
+            ALTER TABLE main_workflowjob ALTER COLUMN char_prompts_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='workflowjob',
+                    name='char_prompts',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='workflowjob',
+            name='char_prompts',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_workflowjob RENAME survey_passwords TO survey_passwords_old;
+            ALTER TABLE main_workflowjob ALTER COLUMN survey_passwords_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='workflowjob',
+                    name='survey_passwords',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='workflowjob',
+            name='survey_passwords',
+            field=models.JSONField(blank=True, default=dict, editable=False),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_workflowjobnode RENAME char_prompts TO char_prompts_old;
+            ALTER TABLE main_workflowjobnode ALTER COLUMN char_prompts_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='workflowjobnode',
+                    name='char_prompts',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='workflowjobnode',
+            name='char_prompts',
+            field=models.JSONField(blank=True, default=dict),
+        ),
+        migrations.RunSQL(
+            """
+            ALTER TABLE main_workflowjobnode RENAME survey_passwords TO survey_passwords_old;
+            ALTER TABLE main_workflowjobnode ALTER COLUMN survey_passwords_old DROP NOT NULL;
+            """,
+            state_operations=[
+                migrations.RemoveField(
+                    model_name='workflowjobnode',
+                    name='survey_passwords',
+                ),
+            ],
+        ),
+        migrations.AddField(
+            model_name='workflowjobnode',
+            name='survey_passwords',
+            field=models.JSONField(blank=True, default=dict, editable=False),
+        ),
+    ]

awx/main/migrations/0186_drop_django_taggit.py

@@ -0,0 +1,27 @@
+# -*- coding: utf-8 -*-
+from __future__ import unicode_literals
+
+from django.db import migrations
+
+
+def delete_taggit_contenttypes(apps, schema_editor):
+    ContentType = apps.get_model('contenttypes', 'ContentType')
+    ContentType.objects.filter(app_label='taggit').delete()
+
+
+def delete_taggit_migration_records(apps, schema_editor):
+    recorder = migrations.recorder.MigrationRecorder(connection=schema_editor.connection)
+    recorder.migration_qs.filter(app='taggit').delete()
+
+
+class Migration(migrations.Migration):
+    dependencies = [
+        ('main', '0185_move_JSONBlob_to_JSONField'),
+    ]
+
+    operations = [
+        migrations.RunSQL("DROP TABLE IF EXISTS taggit_tag CASCADE;"),
+        migrations.RunSQL("DROP TABLE IF EXISTS taggit_taggeditem CASCADE;"),
+        migrations.RunPython(delete_taggit_contenttypes),
+        migrations.RunPython(delete_taggit_migration_records),
+    ]

awx/main/models/__init__.py

@@ -3,6 +3,7 @@
 
 # Django
 from django.conf import settings  # noqa
+from django.db import connection
 from django.db.models.signals import pre_delete  # noqa
 
 # AWX
@@ -99,6 +100,58 @@ User.add_to_class('can_access_with_errors', check_user_access_with_errors)
 User.add_to_class('accessible_objects', user_accessible_objects)
 
 
+def convert_jsonfields():
+    if connection.vendor != 'postgresql':
+        return
+
+    # fmt: off
+    fields = [
+        ('main_activitystream', 'id', (
+            'deleted_actor',
+            'setting',
+        )),
+        ('main_job', 'unifiedjob_ptr_id', (
+            'survey_passwords',
+        )),
+        ('main_joblaunchconfig', 'id', (
+            'char_prompts',
+            'survey_passwords',
+        )),
+        ('main_notification', 'id', (
+            'body',
+        )),
+        ('main_unifiedjob', 'id', (
+            'job_env',
+        )),
+        ('main_workflowjob', 'unifiedjob_ptr_id', (
+            'char_prompts',
+            'survey_passwords',
+        )),
+        ('main_workflowjobnode', 'id', (
+            'char_prompts',
+            'survey_passwords',
+        )),
+    ]
+    # fmt: on
+
+    with connection.cursor() as cursor:
+        for table, pkfield, columns in fields:
+            # Do the renamed old columns still exist?  If so, run the task.
+            old_columns = ','.join(f"'{column}_old'" for column in columns)
+            cursor.execute(
+                f"""
+                select count(1) from information_schema.columns
+                where
+                  table_name = %s and column_name in ({old_columns});
+                """,
+                (table,),
+            )
+            if cursor.fetchone()[0]:
+                from awx.main.tasks.system import migrate_jsonfield
+
+                migrate_jsonfield.apply_async([table, pkfield, columns])
+
+
 def cleanup_created_modified_by(sender, **kwargs):
     # work around a bug in django-polymorphic that doesn't properly
     # handle cascades for reverse foreign keys on the polymorphic base model

awx/main/models/activity_stream.py

@@ -3,7 +3,6 @@
 
 # AWX
 from awx.api.versioning import reverse
-from awx.main.fields import JSONBlob
 from awx.main.models.base import accepts_json
 
 # Django
@@ -36,7 +35,7 @@ class ActivityStream(models.Model):
     operation = models.CharField(max_length=13, choices=OPERATION_CHOICES)
     timestamp = models.DateTimeField(auto_now_add=True)
     changes = accepts_json(models.TextField(blank=True))
-    deleted_actor = JSONBlob(null=True)
+    deleted_actor = models.JSONField(null=True)
     action_node = models.CharField(
         blank=True,
         default='',
@@ -84,7 +83,7 @@ class ActivityStream(models.Model):
     o_auth2_application = models.ManyToManyField("OAuth2Application", blank=True)
     o_auth2_access_token = models.ManyToManyField("OAuth2AccessToken", blank=True)
 
-    setting = JSONBlob(default=dict, blank=True)
+    setting = models.JSONField(default=dict, blank=True)
 
     def __str__(self):
         operation = self.operation if 'operation' in self.__dict__ else '_delayed_'

awx/main/models/base.py

@@ -7,9 +7,6 @@ from django.core.exceptions import ValidationError, ObjectDoesNotExist
 from django.utils.translation import gettext_lazy as _
 from django.utils.timezone import now
 
-# Django-Taggit
-from taggit.managers import TaggableManager
-
 # Django-CRUM
 from crum import get_current_user
 
@@ -301,8 +298,6 @@ class PrimordialModel(HasEditsMixin, CreatedModifiedModel):
         on_delete=models.SET_NULL,
     )
 
-    tags = TaggableManager(blank=True)
-
     def __init__(self, *args, **kwargs):
         r = super(PrimordialModel, self).__init__(*args, **kwargs)
         if self.pk:

awx/main/models/events.py

@@ -4,6 +4,7 @@ import datetime
 from datetime import timezone
 import logging
 from collections import defaultdict
+import time
 
 from django.conf import settings
 from django.core.exceptions import ObjectDoesNotExist
@@ -383,8 +384,17 @@ class BasePlaybookEvent(CreatedModifiedModel):
                         .distinct()
                     )  # noqa
 
-                    job.get_event_queryset().filter(uuid__in=changed).update(changed=True)
-                    job.get_event_queryset().filter(uuid__in=failed).update(failed=True)
+                    # NOTE: we take a set of changed and failed parent uuids because the subquery
+                    # complicates the plan with large event tables causing very long query execution time
+                    changed_start = time.time()
+                    changed_res = job.get_event_queryset().filter(uuid__in=set(changed)).update(changed=True)
+                    failed_start = time.time()
+                    failed_res = job.get_event_queryset().filter(uuid__in=set(failed)).update(failed=True)
+                    logger.debug(
+                        f'Event propagation for job {job.id}: '
+                        f'marked {changed_res} as changed in {failed_start - changed_start:.4f}s, '
+                        f'{failed_res} as failed in {time.time() - failed_start:.4f}s'
+                    )
 
         for field in ('playbook', 'play', 'task', 'role'):
             value = force_str(event_data.get(field, '')).strip()

awx/main/models/ha.py

@@ -20,7 +20,7 @@ from solo.models import SingletonModel
 # AWX
 from awx import __version__ as awx_application_version
 from awx.api.versioning import reverse
-from awx.main.fields import JSONBlob, ImplicitRoleField
+from awx.main.fields import ImplicitRoleField
 from awx.main.managers import InstanceManager, UUID_DEFAULT
 from awx.main.constants import JOB_FOLDER_PREFIX
 from awx.main.models.base import BaseModel, HasEditsMixin, prevent_search
@@ -406,7 +406,7 @@ class InstanceGroup(HasPolicyEditsMixin, BaseModel, RelatedJobsMixin, ResourceMi
     max_forks = models.IntegerField(default=0, help_text=_("Max forks to execute on this group. Zero means no limit."))
     policy_instance_percentage = models.IntegerField(default=0, help_text=_("Percentage of Instances to automatically assign to this group"))
     policy_instance_minimum = models.IntegerField(default=0, help_text=_("Static minimum number of Instances to automatically assign to this group"))
-    policy_instance_list = JSONBlob(
+    policy_instance_list = models.JSONField(
         default=list, blank=True, help_text=_("List of exact-match Instances that will always be automatically assigned to this group")
     )
 

awx/main/models/inventory.py

@@ -899,18 +899,18 @@ class HostMetric(models.Model):
 
             last_automation_before = now() - dateutil.relativedelta.relativedelta(months=months_ago)
 
-            logger.info(f'Cleanup [HostMetric]: soft-deleting records last automated before {last_automation_before}')
+            logger.info(f'cleanup_host_metrics: soft-deleting records last automated before {last_automation_before}')
             HostMetric.active_objects.filter(last_automation__lt=last_automation_before).update(
                 deleted=True, deleted_counter=models.F('deleted_counter') + 1, last_deleted=now()
             )
             settings.CLEANUP_HOST_METRICS_LAST_TS = now()
         except (TypeError, ValueError):
-            logger.error(f"Cleanup [HostMetric]: months_ago({months_ago}) has to be a positive integer value")
+            logger.error(f"cleanup_host_metrics: months_ago({months_ago}) has to be a positive integer value")
 
 
 class HostMetricSummaryMonthly(models.Model):
     """
-    HostMetric summaries computed by scheduled task <TODO> monthly
+    HostMetric summaries computed by scheduled task 'awx.main.tasks.system.host_metric_summary_monthly' monthly
     """
 
     date = models.DateField(unique=True)
@@ -1623,6 +1623,7 @@ class rhv(PluginFileInjector):
     collection = 'ovirt'
     downstream_namespace = 'redhat'
     downstream_collection = 'rhv'
+    use_fqcn = True
 
 
 class satellite6(PluginFileInjector):

awx/main/models/jobs.py

@@ -883,7 +883,7 @@ class LaunchTimeConfigBase(BaseModel):
     )
     # All standard fields are stored in this dictionary field
     # This is a solution to the nullable CharField problem, specific to prompting
-    char_prompts = JSONBlob(default=dict, blank=True)
+    char_prompts = models.JSONField(default=dict, blank=True)
 
     # Define fields that are not really fields, but alias to char_prompts lookups
     limit = NullablePromptPseudoField('limit')
@@ -960,7 +960,7 @@ class LaunchTimeConfig(LaunchTimeConfigBase):
     # Special case prompting fields, even more special than the other ones
     extra_data = JSONBlob(default=dict, blank=True)
     survey_passwords = prevent_search(
-        JSONBlob(
+        models.JSONField(
             default=dict,
             editable=False,
             blank=True,

awx/main/models/mixins.py

@@ -24,7 +24,7 @@ from awx.main.utils import parse_yaml_or_json, get_custom_venv_choices, get_lice
 from awx.main.utils.execution_environments import get_default_execution_environment
 from awx.main.utils.encryption import decrypt_value, get_encryption_key, is_encrypted
 from awx.main.utils.polymorphic import build_polymorphic_ctypes_map
-from awx.main.fields import AskForField, JSONBlob
+from awx.main.fields import AskForField
 from awx.main.constants import ACTIVE_STATES
 
 
@@ -103,7 +103,7 @@ class SurveyJobTemplateMixin(models.Model):
     survey_enabled = models.BooleanField(
         default=False,
     )
-    survey_spec = prevent_search(JSONBlob(default=dict, blank=True))
+    survey_spec = prevent_search(models.JSONField(default=dict, blank=True))
 
     ask_inventory_on_launch = AskForField(
         blank=True,
@@ -392,7 +392,7 @@ class SurveyJobMixin(models.Model):
         abstract = True
 
     survey_passwords = prevent_search(
-        JSONBlob(
+        models.JSONField(
             default=dict,
             editable=False,
             blank=True,

awx/main/models/notifications.py

@@ -17,7 +17,6 @@ from jinja2.exceptions import TemplateSyntaxError, UndefinedError, SecurityError
 
 # AWX
 from awx.api.versioning import reverse
-from awx.main.fields import JSONBlob
 from awx.main.models.base import CommonModelNameNotUnique, CreatedModifiedModel, prevent_search
 from awx.main.utils import encrypt_field, decrypt_field, set_environ
 from awx.main.notifications.email_backend import CustomEmailBackend
@@ -69,12 +68,12 @@ class NotificationTemplate(CommonModelNameNotUnique):
         choices=NOTIFICATION_TYPE_CHOICES,
     )
 
-    notification_configuration = prevent_search(JSONBlob(default=dict))
+    notification_configuration = prevent_search(models.JSONField(default=dict))
 
     def default_messages():
         return {'started': None, 'success': None, 'error': None, 'workflow_approval': None}
 
-    messages = JSONBlob(null=True, blank=True, default=default_messages, help_text=_('Optional custom messages for notification template.'))
+    messages = models.JSONField(null=True, blank=True, default=default_messages, help_text=_('Optional custom messages for notification template.'))
 
     def has_message(self, condition):
         potential_template = self.messages.get(condition, {})
@@ -236,7 +235,7 @@ class Notification(CreatedModifiedModel):
         default='',
         editable=False,
     )
-    body = JSONBlob(default=dict, blank=True)
+    body = models.JSONField(default=dict, blank=True)
 
     def get_absolute_url(self, request=None):
         return reverse('api:notification_detail', kwargs={'pk': self.pk}, request=request)

awx/main/models/projects.py

@@ -33,7 +33,7 @@ from awx.main.models.mixins import ResourceMixin, TaskManagerProjectUpdateMixin,
 from awx.main.utils import update_scm_url, polymorphic
 from awx.main.utils.ansible import skip_directory, could_be_inventory, could_be_playbook
 from awx.main.utils.execution_environments import get_control_plane_execution_environment
-from awx.main.fields import ImplicitRoleField, JSONBlob
+from awx.main.fields import ImplicitRoleField
 from awx.main.models.rbac import (
     ROLE_SINGLETON_SYSTEM_ADMINISTRATOR,
     ROLE_SINGLETON_SYSTEM_AUDITOR,
@@ -303,7 +303,7 @@ class Project(UnifiedJobTemplate, ProjectOptions, ResourceMixin, CustomVirtualEn
         help_text=_('The last revision fetched by a project update'),
     )
 
-    playbook_files = JSONBlob(
+    playbook_files = models.JSONField(
         default=list,
         blank=True,
         editable=False,
@@ -311,7 +311,7 @@ class Project(UnifiedJobTemplate, ProjectOptions, ResourceMixin, CustomVirtualEn
         help_text=_('List of playbooks found in the project'),
     )
 
-    inventory_files = JSONBlob(
+    inventory_files = models.JSONField(
         default=list,
         blank=True,
         editable=False,
@@ -479,7 +479,7 @@ class Project(UnifiedJobTemplate, ProjectOptions, ResourceMixin, CustomVirtualEn
         RunProjectUpdate/RunInventoryUpdate.
         """
 
-        if self.status not in ('error', 'failed'):
+        if self.status not in ('error', 'failed') or self.scm_update_on_launch:
             return None
 
         latest_update = self.project_updates.last()

awx/main/models/unified_jobs.py

@@ -55,7 +55,7 @@ from awx.main.utils import polymorphic
 from awx.main.constants import ACTIVE_STATES, CAN_CANCEL, JOB_VARIABLE_PREFIXES
 from awx.main.redact import UriCleaner, REPLACE_STR
 from awx.main.consumers import emit_channel_notification
-from awx.main.fields import AskForField, OrderedManyToManyField, JSONBlob
+from awx.main.fields import AskForField, OrderedManyToManyField
 
 __all__ = ['UnifiedJobTemplate', 'UnifiedJob', 'StdoutMaxBytesExceeded']
 
@@ -668,7 +668,7 @@ class UnifiedJob(
         editable=False,
     )
     job_env = prevent_search(
-        JSONBlob(
+        models.JSONField(
             default=dict,
             blank=True,
             editable=False,
@@ -1137,11 +1137,6 @@ class UnifiedJob(
                     if total > max_supported:
                         raise StdoutMaxBytesExceeded(total, max_supported)
 
-                # psycopg2's copy_expert writes bytes, but callers of this
-                # function assume a str-based fd will be returned; decode
-                # .write() calls on the fly to maintain this interface
-                _write = fd.write
-                fd.write = lambda s: _write(smart_str(s))
                 tbl = self._meta.db_table + 'event'
                 created_by_cond = ''
                 if self.has_unpartitioned_events:
@@ -1150,7 +1145,12 @@ class UnifiedJob(
                     created_by_cond = f"job_created='{self.created.isoformat()}' AND "
 
                 sql = f"copy (select stdout from {tbl} where {created_by_cond}{self.event_parent_key}={self.id} and stdout != '' order by start_line) to stdout"  # nosql
-                cursor.copy_expert(sql, fd)
+                # psycopg3's copy writes bytes, but callers of this
+                # function assume a str-based fd will be returned; decode
+                # .write() calls on the fly to maintain this interface
+                with cursor.copy(sql) as copy:
+                    while data := copy.read():
+                        fd.write(smart_str(bytes(data)))
 
                 if hasattr(fd, 'name'):
                     # If we're dealing with a physical file, use `sed` to clean

awx/main/models/workflow.py

@@ -661,7 +661,11 @@ class WorkflowJob(UnifiedJob, WorkflowJobOptions, SurveyJobMixin, JobNotificatio
 
     @property
     def event_processing_finished(self):
-        return True
+        return True  # workflow jobs do not have events
+
+    @property
+    def has_unpartitioned_events(self):
+        return False  # workflow jobs do not have events
 
     def _get_parent_field_name(self):
         if self.job_template_id:
@@ -914,7 +918,11 @@ class WorkflowApproval(UnifiedJob, JobNotificationMixin):
 
     @property
     def event_processing_finished(self):
-        return True
+        return True  # approval jobs do not have events
+
+    @property
+    def has_unpartitioned_events(self):
+        return False  # approval jobs do not have events
 
     def send_approval_notification(self, approval_status):
         from awx.main.tasks.system import send_notifications  # avoid circular import

awx/main/registrar.py

@@ -3,8 +3,6 @@
 
 from django.db.models.signals import pre_save, post_save, pre_delete, m2m_changed
 
-from taggit.managers import TaggableManager
-
 
 class ActivityStreamRegistrar(object):
     def __init__(self):
@@ -21,8 +19,6 @@ class ActivityStreamRegistrar(object):
             pre_delete.connect(activity_stream_delete, sender=model, dispatch_uid=str(self.__class__) + str(model) + "_delete")
 
             for m2mfield in model._meta.many_to_many:
-                if isinstance(m2mfield, TaggableManager):
-                    continue  # Special case for taggit app
                 try:
                     m2m_attr = getattr(model, m2mfield.name)
                     m2m_changed.connect(

awx/main/scheduler/task_manager.py

@@ -25,7 +25,6 @@ from awx.main.models import (
     InventoryUpdate,
     Job,
     Project,
-    ProjectUpdate,
     UnifiedJob,
     WorkflowApproval,
     WorkflowJob,
@@ -102,27 +101,33 @@ class TaskBase:
 
     def record_aggregate_metrics(self, *args):
         if not is_testing():
-            # increment task_manager_schedule_calls regardless if the other
-            # metrics are recorded
-            s_metrics.Metrics(auto_pipe_execute=True).inc(f"{self.prefix}__schedule_calls", 1)
-            # Only record metrics if the last time recording was more
-            # than SUBSYSTEM_METRICS_TASK_MANAGER_RECORD_INTERVAL ago.
-            # Prevents a short-duration task manager that runs directly after a
-            # long task manager to override useful metrics.
-            current_time = time.time()
-            time_last_recorded = current_time - self.subsystem_metrics.decode(f"{self.prefix}_recorded_timestamp")
-            if time_last_recorded > settings.SUBSYSTEM_METRICS_TASK_MANAGER_RECORD_INTERVAL:
-                logger.debug(f"recording {self.prefix} metrics, last recorded {time_last_recorded} seconds ago")
-                self.subsystem_metrics.set(f"{self.prefix}_recorded_timestamp", current_time)
-                self.subsystem_metrics.pipe_execute()
-            else:
-                logger.debug(f"skipping recording {self.prefix} metrics, last recorded {time_last_recorded} seconds ago")
+            try:
+                # increment task_manager_schedule_calls regardless if the other
+                # metrics are recorded
+                s_metrics.Metrics(auto_pipe_execute=True).inc(f"{self.prefix}__schedule_calls", 1)
+                # Only record metrics if the last time recording was more
+                # than SUBSYSTEM_METRICS_TASK_MANAGER_RECORD_INTERVAL ago.
+                # Prevents a short-duration task manager that runs directly after a
+                # long task manager to override useful metrics.
+                current_time = time.time()
+                time_last_recorded = current_time - self.subsystem_metrics.decode(f"{self.prefix}_recorded_timestamp")
+                if time_last_recorded > settings.SUBSYSTEM_METRICS_TASK_MANAGER_RECORD_INTERVAL:
+                    logger.debug(f"recording {self.prefix} metrics, last recorded {time_last_recorded} seconds ago")
+                    self.subsystem_metrics.set(f"{self.prefix}_recorded_timestamp", current_time)
+                    self.subsystem_metrics.pipe_execute()
+                else:
+                    logger.debug(f"skipping recording {self.prefix} metrics, last recorded {time_last_recorded} seconds ago")
+            except Exception:
+                logger.exception(f"Error saving metrics for {self.prefix}")
 
     def record_aggregate_metrics_and_exit(self, *args):
         self.record_aggregate_metrics()
         sys.exit(1)
 
     def schedule(self):
+        # Always be able to restore the original signal handler if we finish
+        original_sigusr1 = signal.getsignal(signal.SIGUSR1)
+
         # Lock
         with task_manager_bulk_reschedule():
             with advisory_lock(f"{self.prefix}_lock", wait=False) as acquired:
@@ -131,9 +136,14 @@ class TaskBase:
                         logger.debug(f"Not running {self.prefix} scheduler, another task holds lock")
                         return
                     logger.debug(f"Starting {self.prefix} Scheduler")
-                    # if sigterm due to timeout, still record metrics
-                    signal.signal(signal.SIGTERM, self.record_aggregate_metrics_and_exit)
-                    self._schedule()
+                    # if sigusr1 due to timeout, still record metrics
+                    signal.signal(signal.SIGUSR1, self.record_aggregate_metrics_and_exit)
+                    try:
+                        self._schedule()
+                    finally:
+                        # Reset the signal handler back to the default just in case anything
+                        # else uses the same signal for other purposes
+                        signal.signal(signal.SIGUSR1, original_sigusr1)
                     commit_start = time.time()
 
                 if self.prefix == "task_manager":
@@ -154,7 +164,6 @@ class WorkflowManager(TaskBase):
                 logger.warning("Workflow manager has reached time out while processing running workflows, exiting loop early")
                 ScheduleWorkflowManager().schedule()
                 # Do not process any more workflow jobs. Stop here.
-                # Maybe we should schedule another WorkflowManager run
                 break
             dag = WorkflowDAG(workflow_job)
             status_changed = False
@@ -169,8 +178,8 @@ class WorkflowManager(TaskBase):
                     workflow_job.save(update_fields=['status', 'start_args'])
                     status_changed = True
             else:
-                workflow_nodes = dag.mark_dnr_nodes()
-                WorkflowJobNode.objects.bulk_update(workflow_nodes, ['do_not_run'])
+                dnr_nodes = dag.mark_dnr_nodes()
+                WorkflowJobNode.objects.bulk_update(dnr_nodes, ['do_not_run'])
                 # If workflow is now done, we do special things to mark it as done.
                 is_done = dag.is_workflow_done()
                 if is_done:
@@ -250,6 +259,7 @@ class WorkflowManager(TaskBase):
                         job.status = 'failed'
                         job.save(update_fields=['status', 'job_explanation'])
                         job.websocket_emit_status('failed')
+                        ScheduleWorkflowManager().schedule()
 
                     # TODO: should we emit a status on the socket here similar to tasks.py awx_periodic_scheduler() ?
                     # emit_websocket_notification('/socket.io/jobs', '', dict(id=))
@@ -270,184 +280,115 @@ class WorkflowManager(TaskBase):
 class DependencyManager(TaskBase):
     def __init__(self):
         super().__init__(prefix="dependency_manager")
+        self.all_projects = {}
+        self.all_inventory_sources = {}
 
-    def create_project_update(self, task, project_id=None):
-        if project_id is None:
-            project_id = task.project_id
-        project_task = Project.objects.get(id=project_id).create_project_update(_eager_fields=dict(launch_type='dependency'))
-
-        # Project created 1 seconds behind
-        project_task.created = task.created - timedelta(seconds=1)
-        project_task.status = 'pending'
-        project_task.save()
-        logger.debug('Spawned {} as dependency of {}'.format(project_task.log_format, task.log_format))
-        return project_task
-
-    def create_inventory_update(self, task, inventory_source_task):
-        inventory_task = InventorySource.objects.get(id=inventory_source_task.id).create_inventory_update(_eager_fields=dict(launch_type='dependency'))
-
-        inventory_task.created = task.created - timedelta(seconds=2)
-        inventory_task.status = 'pending'
-        inventory_task.save()
-        logger.debug('Spawned {} as dependency of {}'.format(inventory_task.log_format, task.log_format))
-
-        return inventory_task
-
-    def add_dependencies(self, task, dependencies):
-        with disable_activity_stream():
-            task.dependent_jobs.add(*dependencies)
-
-    def get_inventory_source_tasks(self):
+    def cache_projects_and_sources(self, task_list):
+        project_ids = set()
         inventory_ids = set()
-        for task in self.all_tasks:
+        for task in task_list:
             if isinstance(task, Job):
-                inventory_ids.add(task.inventory_id)
-        self.all_inventory_sources = [invsrc for invsrc in InventorySource.objects.filter(inventory_id__in=inventory_ids, update_on_launch=True)]
+                if task.project_id:
+                    project_ids.add(task.project_id)
+                if task.inventory_id:
+                    inventory_ids.add(task.inventory_id)
+            elif isinstance(task, InventoryUpdate):
+                if task.inventory_source and task.inventory_source.source_project_id:
+                    project_ids.add(task.inventory_source.source_project_id)
 
-    def get_latest_inventory_update(self, inventory_source):
-        latest_inventory_update = InventoryUpdate.objects.filter(inventory_source=inventory_source).order_by("-created")
-        if not latest_inventory_update.exists():
-            return None
-        return latest_inventory_update.first()
+        for proj in Project.objects.filter(id__in=project_ids, scm_update_on_launch=True):
+            self.all_projects[proj.id] = proj
 
-    def should_update_inventory_source(self, job, latest_inventory_update):
-        now = tz_now()
+        for invsrc in InventorySource.objects.filter(inventory_id__in=inventory_ids, update_on_launch=True):
+            self.all_inventory_sources.setdefault(invsrc.inventory_id, [])
+            self.all_inventory_sources[invsrc.inventory_id].append(invsrc)
 
-        if latest_inventory_update is None:
+    @staticmethod
+    def should_update_again(update, cache_timeout):
+        '''
+        If it has never updated, we need to update
+        If there is already an update in progress then we do not need to a new create one
+        If the last update failed, we always need to try and update again
+        If current time is more than cache_timeout after last update, then we need a new one
+        '''
+        if (update is None) or (update.status in ['failed', 'canceled', 'error']):
             return True
-        '''
-        If there's already a inventory update utilizing this job that's about to run
-        then we don't need to create one
-        '''
-        if latest_inventory_update.status in ['waiting', 'pending', 'running']:
+        if update.status in ['waiting', 'pending', 'running']:
             return False
 
-        timeout_seconds = timedelta(seconds=latest_inventory_update.inventory_source.update_cache_timeout)
-        if (latest_inventory_update.finished + timeout_seconds) < now:
-            return True
-        if latest_inventory_update.inventory_source.update_on_launch is True and latest_inventory_update.status in ['failed', 'canceled', 'error']:
-            return True
-        return False
+        return bool(((update.finished + timedelta(seconds=cache_timeout))) < tz_now())
 
-    def get_latest_project_update(self, project_id):
-        latest_project_update = ProjectUpdate.objects.filter(project=project_id, job_type='check').order_by("-created")
-        if not latest_project_update.exists():
-            return None
-        return latest_project_update.first()
-
-    def should_update_related_project(self, job, latest_project_update):
-        now = tz_now()
-
-        if latest_project_update is None:
-            return True
-
-        if latest_project_update.status in ['failed', 'canceled']:
-            return True
-
-        '''
-        If there's already a project update utilizing this job that's about to run
-        then we don't need to create one
-        '''
-        if latest_project_update.status in ['waiting', 'pending', 'running']:
-            return False
-
-        '''
-        If the latest project update has a created time == job_created_time-1
-        then consider the project update found. This is so we don't enter an infinite loop
-        of updating the project when cache timeout is 0.
-        '''
-        if (
-            latest_project_update.project.scm_update_cache_timeout == 0
-            and latest_project_update.launch_type == 'dependency'
-            and latest_project_update.created == job.created - timedelta(seconds=1)
-        ):
-            return False
-        '''
-        Normal Cache Timeout Logic
-        '''
-        timeout_seconds = timedelta(seconds=latest_project_update.project.scm_update_cache_timeout)
-        if (latest_project_update.finished + timeout_seconds) < now:
-            return True
-        return False
+    def get_or_create_project_update(self, project_id):
+        project = self.all_projects.get(project_id, None)
+        if project is not None:
+            latest_project_update = project.project_updates.filter(job_type='check').order_by("-created").first()
+            if self.should_update_again(latest_project_update, project.scm_update_cache_timeout):
+                project_task = project.create_project_update(_eager_fields=dict(launch_type='dependency'))
+                project_task.signal_start()
+                return [project_task]
+            else:
+                return [latest_project_update]
+        return []
 
     def gen_dep_for_job(self, task):
-        created_dependencies = []
-        dependencies = []
-        # TODO: Can remove task.project None check after scan-job-default-playbook is removed
-        if task.project is not None and task.project.scm_update_on_launch is True:
-            latest_project_update = self.get_latest_project_update(task.project_id)
-            if self.should_update_related_project(task, latest_project_update):
-                latest_project_update = self.create_project_update(task)
-                created_dependencies.append(latest_project_update)
-            dependencies.append(latest_project_update)
+        dependencies = self.get_or_create_project_update(task.project_id)
 
-        # Inventory created 2 seconds behind job
         try:
             start_args = json.loads(decrypt_field(task, field_name="start_args"))
         except ValueError:
             start_args = dict()
-        # generator for inventory sources related to this task
-        task_inv_sources = (invsrc for invsrc in self.all_inventory_sources if invsrc.inventory_id == task.inventory_id)
-        for inventory_source in task_inv_sources:
+        # generator for update-on-launch inventory sources related to this task
+        for inventory_source in self.all_inventory_sources.get(task.inventory_id, []):
             if "inventory_sources_already_updated" in start_args and inventory_source.id in start_args['inventory_sources_already_updated']:
                 continue
-            if not inventory_source.update_on_launch:
-                continue
-            latest_inventory_update = self.get_latest_inventory_update(inventory_source)
-            if self.should_update_inventory_source(task, latest_inventory_update):
-                inventory_task = self.create_inventory_update(task, inventory_source)
-                created_dependencies.append(inventory_task)
+            latest_inventory_update = inventory_source.inventory_updates.order_by("-created").first()
+            if self.should_update_again(latest_inventory_update, inventory_source.update_cache_timeout):
+                inventory_task = inventory_source.create_inventory_update(_eager_fields=dict(launch_type='dependency'))
+                inventory_task.signal_start()
                 dependencies.append(inventory_task)
             else:
                 dependencies.append(latest_inventory_update)
 
-        if dependencies:
-            self.add_dependencies(task, dependencies)
-
-        return created_dependencies
+        return dependencies
 
     def gen_dep_for_inventory_update(self, inventory_task):
-        created_dependencies = []
         if inventory_task.source == "scm":
             invsrc = inventory_task.inventory_source
-            if not invsrc.source_project.scm_update_on_launch:
-                return created_dependencies
-
-            latest_src_project_update = self.get_latest_project_update(invsrc.source_project_id)
-            if self.should_update_related_project(inventory_task, latest_src_project_update):
-                latest_src_project_update = self.create_project_update(inventory_task, project_id=invsrc.source_project_id)
-                created_dependencies.append(latest_src_project_update)
-            self.add_dependencies(inventory_task, [latest_src_project_update])
-            latest_src_project_update.scm_inventory_updates.add(inventory_task)
-        return created_dependencies
+            if invsrc:
+                return self.get_or_create_project_update(invsrc.source_project_id)
+        return []
 
     @timeit
     def generate_dependencies(self, undeped_tasks):
-        created_dependencies = []
+        dependencies = []
+        self.cache_projects_and_sources(undeped_tasks)
         for task in undeped_tasks:
             task.log_lifecycle("acknowledged")
             if type(task) is Job:
-                created_dependencies += self.gen_dep_for_job(task)
+                job_deps = self.gen_dep_for_job(task)
             elif type(task) is InventoryUpdate:
-                created_dependencies += self.gen_dep_for_inventory_update(task)
+                job_deps = self.gen_dep_for_inventory_update(task)
             else:
                 continue
+            if job_deps:
+                dependencies += job_deps
+                with disable_activity_stream():
+                    task.dependent_jobs.add(*dependencies)
+                logger.debug(f'Linked {[dep.log_format for dep in dependencies]} as dependencies of {task.log_format}')
+
         UnifiedJob.objects.filter(pk__in=[task.pk for task in undeped_tasks]).update(dependencies_processed=True)
 
-        return created_dependencies
-
-    def process_tasks(self):
-        deps = self.generate_dependencies(self.all_tasks)
-        self.generate_dependencies(deps)
-        self.subsystem_metrics.inc(f"{self.prefix}_pending_processed", len(self.all_tasks) + len(deps))
+        return dependencies
 
     @timeit
     def _schedule(self):
         self.get_tasks(dict(status__in=["pending"], dependencies_processed=False))
 
         if len(self.all_tasks) > 0:
-            self.get_inventory_source_tasks()
-            self.process_tasks()
+            deps = self.generate_dependencies(self.all_tasks)
+            undeped_deps = [dep for dep in deps if dep.dependencies_processed is False]
+            self.generate_dependencies(undeped_deps)
+            self.subsystem_metrics.inc(f"{self.prefix}_pending_processed", len(self.all_tasks) + len(undeped_deps))
             ScheduleTaskManager().schedule()
 
 

awx/main/tasks/__init__.py

@@ -1 +1 @@
-from . import jobs, receptor, system  # noqa
+from . import host_metrics, jobs, receptor, system  # noqa

awx/main/tasks/callback.py

@@ -31,6 +31,7 @@ class RunnerCallback:
         self.model = model
         self.update_attempts = int(settings.DISPATCHER_DB_DOWNTOWN_TOLLERANCE / 5)
         self.wrapup_event_dispatched = False
+        self.artifacts_processed = False
         self.extra_update_fields = {}
 
     def update_model(self, pk, _attempt=0, **updates):
@@ -211,6 +212,9 @@ class RunnerCallback:
             if result_traceback:
                 self.delay_update(result_traceback=result_traceback)
 
+    def artifacts_handler(self, artifact_dir):
+        self.artifacts_processed = True
+
 
 class RunnerCallbackForProjectUpdate(RunnerCallback):
     def __init__(self, *args, **kwargs):

awx/main/tasks/facts.py

@@ -9,6 +9,7 @@ from django.conf import settings
 from django.db.models.query import QuerySet
 from django.utils.encoding import smart_str
 from django.utils.timezone import now
+from django.db import OperationalError
 
 # AWX
 from awx.main.utils.common import log_excess_runtime
@@ -57,6 +58,28 @@ def start_fact_cache(hosts, destination, log_data, timeout=None, inventory_id=No
     return None
 
 
+def raw_update_hosts(host_list):
+    Host.objects.bulk_update(host_list, ['ansible_facts', 'ansible_facts_modified'])
+
+
+def update_hosts(host_list, max_tries=5):
+    if not host_list:
+        return
+    for i in range(max_tries):
+        try:
+            raw_update_hosts(host_list)
+        except OperationalError as exc:
+            # Deadlocks can happen if this runs at the same time as another large query
+            # inventory updates and updating last_job_host_summary are candidates for conflict
+            # but these would resolve easily on a retry
+            if i + 1 < max_tries:
+                logger.info(f'OperationalError (suspected deadlock) saving host facts retry {i}, message: {exc}')
+                continue
+            else:
+                raise
+        break
+
+
 @log_excess_runtime(
     logger,
     debug_cutoff=0.01,
@@ -111,7 +134,6 @@ def finish_fact_cache(hosts, destination, facts_write_time, log_data, job_id=Non
             system_tracking_logger.info('Facts cleared for inventory {} host {}'.format(smart_str(host.inventory.name), smart_str(host.name)))
             log_data['cleared_ct'] += 1
         if len(hosts_to_update) > 100:
-            Host.objects.bulk_update(hosts_to_update, ['ansible_facts', 'ansible_facts_modified'])
+            update_hosts(hosts_to_update)
             hosts_to_update = []
-    if hosts_to_update:
-        Host.objects.bulk_update(hosts_to_update, ['ansible_facts', 'ansible_facts_modified'])
+    update_hosts(hosts_to_update)

awx/main/tasks/host_metrics.py

@@ -0,0 +1,205 @@
+import datetime
+from dateutil.relativedelta import relativedelta
+import logging
+
+from django.conf import settings
+from django.db.models import Count
+from django.db.models.functions import TruncMonth
+from django.utils.timezone import now
+from rest_framework.fields import DateTimeField
+from awx.main.dispatch import get_task_queuename
+from awx.main.dispatch.publish import task
+from awx.main.models.inventory import HostMetric, HostMetricSummaryMonthly
+from awx.conf.license import get_license
+
+logger = logging.getLogger('awx.main.tasks.host_metric_summary_monthly')
+
+
+@task(queue=get_task_queuename)
+def host_metric_summary_monthly():
+    """Run cleanup host metrics summary monthly task each week"""
+    if _is_run_threshold_reached(
+        getattr(settings, 'HOST_METRIC_SUMMARY_TASK_LAST_TS', None), getattr(settings, 'HOST_METRIC_SUMMARY_TASK_INTERVAL', 7) * 86400
+    ):
+        logger.info(f"Executing host_metric_summary_monthly, last ran at {getattr(settings, 'HOST_METRIC_SUMMARY_TASK_LAST_TS', '---')}")
+        HostMetricSummaryMonthlyTask().execute()
+        logger.info("Finished host_metric_summary_monthly")
+
+
+def _is_run_threshold_reached(setting, threshold_seconds):
+    last_time = DateTimeField().to_internal_value(setting) if setting else DateTimeField().to_internal_value('1970-01-01')
+
+    return (now() - last_time).total_seconds() > threshold_seconds
+
+
+class HostMetricSummaryMonthlyTask:
+    """
+    This task computes last [threshold] months of HostMetricSummaryMonthly table
+    [threshold] is setting CLEANUP_HOST_METRICS_HARD_THRESHOLD
+    Each record in the table represents changes in HostMetric table in one month
+    It always overrides all the months newer than <threshold>, never updates older months
+    Algorithm:
+    - hosts_added are HostMetric records with first_automation in given month
+    - hosts_deleted are HostMetric records with deleted=True and last_deleted in given month
+    - - HostMetrics soft-deleted before <threshold> also increases hosts_deleted in their last_deleted month
+    - license_consumed is license_consumed(previous month) + hosts_added - hosts_deleted
+    - - license_consumed for HostMetricSummaryMonthly.date < [threshold] is computed also from
+        all HostMetrics.first_automation < [threshold]
+    - license_capacity is set only for current month, and it's never updated (value taken from current subscription)
+    """
+
+    def __init__(self):
+        self.host_metrics = {}
+        self.processed_month = self._get_first_month()
+        self.existing_summaries = None
+        self.existing_summaries_idx = 0
+        self.existing_summaries_cnt = 0
+        self.records_to_create = []
+        self.records_to_update = []
+
+    def execute(self):
+        self._load_existing_summaries()
+        self._load_hosts_added()
+        self._load_hosts_deleted()
+
+        # Get first month after last hard delete
+        month = self._get_first_month()
+        license_consumed = self._get_license_consumed_before(month)
+
+        # Fill record for each month
+        while month <= datetime.date.today().replace(day=1):
+            summary = self._find_or_create_summary(month)
+            # Update summary and update license_consumed by hosts added/removed this month
+            self._update_summary(summary, month, license_consumed)
+            license_consumed = summary.license_consumed
+
+            month = month + relativedelta(months=1)
+
+        # Create/Update stats
+        HostMetricSummaryMonthly.objects.bulk_create(self.records_to_create, batch_size=1000)
+        HostMetricSummaryMonthly.objects.bulk_update(self.records_to_update, ['license_consumed', 'hosts_added', 'hosts_deleted'], batch_size=1000)
+
+        # Set timestamp of last run
+        settings.HOST_METRIC_SUMMARY_TASK_LAST_TS = now()
+
+    def _get_license_consumed_before(self, month):
+        license_consumed = 0
+        for metric_month, metric in self.host_metrics.items():
+            if metric_month < month:
+                hosts_added = metric.get('hosts_added', 0)
+                hosts_deleted = metric.get('hosts_deleted', 0)
+                license_consumed = license_consumed + hosts_added - hosts_deleted
+            else:
+                break
+        return license_consumed
+
+    def _load_existing_summaries(self):
+        """Find all summaries newer than host metrics delete threshold"""
+        self.existing_summaries = HostMetricSummaryMonthly.objects.filter(date__gte=self._get_first_month()).order_by('date')
+        self.existing_summaries_idx = 0
+        self.existing_summaries_cnt = len(self.existing_summaries)
+
+    def _load_hosts_added(self):
+        """Aggregates hosts added each month, by the 'first_automation' timestamp"""
+        #
+        # -- SQL translation (for better code readability)
+        # SELECT date_trunc('month', first_automation) as month,
+        #        count(first_automation) AS hosts_added
+        # FROM main_hostmetric
+        # GROUP BY month
+        # ORDER by month;
+        result = (
+            HostMetric.objects.annotate(month=TruncMonth('first_automation'))
+            .values('month')
+            .annotate(hosts_added=Count('first_automation'))
+            .values('month', 'hosts_added')
+            .order_by('month')
+        )
+
+        for host_metric in list(result):
+            month = host_metric['month']
+            if month:
+                beginning_of_month = datetime.date(month.year, month.month, 1)
+                if self.host_metrics.get(beginning_of_month) is None:
+                    self.host_metrics[beginning_of_month] = {}
+                self.host_metrics[beginning_of_month]['hosts_added'] = host_metric['hosts_added']
+
+    def _load_hosts_deleted(self):
+        """
+        Aggregates hosts deleted each month, by the 'last_deleted' timestamp.
+        Host metrics have to be deleted NOW to be counted as deleted before
+        (by intention - statistics can change retrospectively by re-automation of previously deleted host)
+        """
+        #
+        # -- SQL translation (for better code readability)
+        # SELECT date_trunc('month', last_deleted) as month,
+        #       count(last_deleted) AS hosts_deleted
+        # FROM main_hostmetric
+        # WHERE deleted = True
+        # GROUP BY 1 # equal to "GROUP BY month"
+        # ORDER by month;
+        result = (
+            HostMetric.objects.annotate(month=TruncMonth('last_deleted'))
+            .values('month')
+            .annotate(hosts_deleted=Count('last_deleted'))
+            .values('month', 'hosts_deleted')
+            .filter(deleted=True)
+            .order_by('month')
+        )
+        for host_metric in list(result):
+            month = host_metric['month']
+            if month:
+                beginning_of_month = datetime.date(month.year, month.month, 1)
+                if self.host_metrics.get(beginning_of_month) is None:
+                    self.host_metrics[beginning_of_month] = {}
+                self.host_metrics[beginning_of_month]['hosts_deleted'] = host_metric['hosts_deleted']
+
+    def _find_or_create_summary(self, month):
+        summary = self._find_summary(month)
+
+        if not summary:
+            summary = HostMetricSummaryMonthly(date=month)
+            self.records_to_create.append(summary)
+        else:
+            self.records_to_update.append(summary)
+        return summary
+
+    def _find_summary(self, month):
+        """
+        Existing summaries are ordered by month ASC.
+        This method is called with month in ascending order too => only 1 traversing is enough
+        """
+        summary = None
+        while not summary and self.existing_summaries_idx < self.existing_summaries_cnt:
+            tmp = self.existing_summaries[self.existing_summaries_idx]
+            if tmp.date < month:
+                self.existing_summaries_idx += 1
+            elif tmp.date == month:
+                summary = tmp
+            elif tmp.date > month:
+                break
+        return summary
+
+    def _update_summary(self, summary, month, license_consumed):
+        """Updates the metric with hosts added and deleted and set license info for current month"""
+        # Get month counts from host metrics, zero if not found
+        hosts_added, hosts_deleted = 0, 0
+        if metric := self.host_metrics.get(month, None):
+            hosts_added = metric.get('hosts_added', 0)
+            hosts_deleted = metric.get('hosts_deleted', 0)
+
+        summary.license_consumed = license_consumed + hosts_added - hosts_deleted
+        summary.hosts_added = hosts_added
+        summary.hosts_deleted = hosts_deleted
+
+        # Set subscription count for current month
+        if month == datetime.date.today().replace(day=1):
+            license_info = get_license()
+            summary.license_capacity = license_info.get('instance_count', 0)
+        return summary
+
+    @staticmethod
+    def _get_first_month():
+        """Returns first month after host metrics hard delete threshold"""
+        threshold = getattr(settings, 'CLEANUP_HOST_METRICS_HARD_THRESHOLD', 36)
+        return datetime.date.today().replace(day=1) - relativedelta(months=int(threshold) - 1)

awx/main/tasks/jobs.py

@@ -290,13 +290,6 @@ class BaseTask(object):
             content = safe_dump(vars, safe_dict)
         return self.write_private_data_file(private_data_dir, 'extravars', content, sub_dir='env')
 
-    def add_awx_venv(self, env):
-        env['VIRTUAL_ENV'] = settings.AWX_VENV_PATH
-        if 'PATH' in env:
-            env['PATH'] = os.path.join(settings.AWX_VENV_PATH, "bin") + ":" + env['PATH']
-        else:
-            env['PATH'] = os.path.join(settings.AWX_VENV_PATH, "bin")
-
     def build_env(self, instance, private_data_dir, private_data_files=None):
         """
         Build environment dictionary for ansible-playbook.
@@ -926,6 +919,7 @@ class RunJob(SourceControlMixin, BaseTask):
         path_vars = (
             ('ANSIBLE_COLLECTIONS_PATHS', 'collections_paths', 'requirements_collections', '~/.ansible/collections:/usr/share/ansible/collections'),
             ('ANSIBLE_ROLES_PATH', 'roles_path', 'requirements_roles', '~/.ansible/roles:/usr/share/ansible/roles:/etc/ansible/roles'),
+            ('ANSIBLE_COLLECTIONS_PATH', 'collections_path', 'requirements_collections', '~/.ansible/collections:/usr/share/ansible/collections'),
         )
 
         config_values = read_ansible_config(os.path.join(private_data_dir, 'project'), list(map(lambda x: x[1], path_vars)))
@@ -1100,7 +1094,7 @@ class RunJob(SourceControlMixin, BaseTask):
             # actual `run()` call; this _usually_ means something failed in
             # the pre_run_hook method
             return
-        if self.should_use_fact_cache():
+        if self.should_use_fact_cache() and self.runner_callback.artifacts_processed:
             job.log_lifecycle("finish_job_fact_cache")
             finish_fact_cache(
                 job.get_hosts_for_fact_cache(),
@@ -1268,7 +1262,7 @@ class RunProjectUpdate(BaseTask):
 
         galaxy_creds_are_defined = project_update.project.organization and project_update.project.organization.galaxy_credentials.exists()
         if not galaxy_creds_are_defined and (settings.AWX_ROLES_ENABLED or settings.AWX_COLLECTIONS_ENABLED):
-            logger.warning('Galaxy role/collection syncing is enabled, but no ' f'credentials are configured for {project_update.project.organization}.')
+            logger.warning('Galaxy role/collection syncing is enabled, but no credentials are configured for {project_update.project.organization}.')
 
         extra_vars.update(
             {

awx/main/tasks/receptor.py

@@ -464,6 +464,7 @@ class AWXReceptorJob:
             event_handler=self.task.runner_callback.event_handler,
             finished_callback=self.task.runner_callback.finished_callback,
             status_handler=self.task.runner_callback.status_handler,
+            artifacts_handler=self.task.runner_callback.artifacts_handler,
             **self.runner_params,
         )
 
@@ -639,11 +640,11 @@ class AWXReceptorJob:
 #
 RECEPTOR_CONFIG_STARTER = (
     {'local-only': None},
-    {'log-level': 'info'},
+    {'log-level': settings.RECEPTOR_LOG_LEVEL},
     {'node': {'firewallrules': [{'action': 'reject', 'tonode': settings.CLUSTER_HOST_ID, 'toservice': 'control'}]}},
     {'control-service': {'service': 'control', 'filename': '/var/run/receptor/receptor.sock', 'permissions': '0660'}},
     {'work-command': {'worktype': 'local', 'command': 'ansible-runner', 'params': 'worker', 'allowruntimeparams': True}},
-    {'work-signing': {'privatekey': '/etc/receptor/signing/work-private-key.pem', 'tokenexpiration': '1m'}},
+    {'work-signing': {'privatekey': '/etc/receptor/work_private_key.pem', 'tokenexpiration': '1m'}},
     {
         'work-kubernetes': {
             'worktype': 'kubernetes-runtime-auth',
@@ -665,7 +666,7 @@ RECEPTOR_CONFIG_STARTER = (
     {
         'tls-client': {
             'name': 'tlsclient',
-            'rootcas': '/etc/receptor/tls/ca/receptor-ca.crt',
+            'rootcas': '/etc/receptor/tls/ca/mesh-CA.crt',
             'cert': '/etc/receptor/tls/receptor.crt',
             'key': '/etc/receptor/tls/receptor.key',
             'mintls13': False,

awx/main/tasks/system.py

@@ -2,6 +2,7 @@
 from collections import namedtuple
 import functools
 import importlib
+import itertools
 import json
 import logging
 import os
@@ -14,7 +15,7 @@ from datetime import datetime
 
 # Django
 from django.conf import settings
-from django.db import transaction, DatabaseError, IntegrityError
+from django.db import connection, transaction, DatabaseError, IntegrityError
 from django.db.models.fields.related import ForeignKey
 from django.utils.timezone import now, timedelta
 from django.utils.encoding import smart_str
@@ -48,6 +49,7 @@ from awx.main.models import (
     SmartInventoryMembership,
     Job,
     HostMetric,
+    convert_jsonfields,
 )
 from awx.main.constants import ACTIVE_STATES
 from awx.main.dispatch.publish import task
@@ -86,6 +88,11 @@ def dispatch_startup():
     if settings.IS_K8S:
         write_receptor_config()
 
+    try:
+        convert_jsonfields()
+    except Exception:
+        logger.exception("Failed json field conversion, skipping.")
+
     startup_logger.debug("Syncing Schedules")
     for sch in Schedule.objects.all():
         try:
@@ -129,6 +136,52 @@ def inform_cluster_of_shutdown():
         logger.exception('Encountered problem with normal shutdown signal.')
 
 
+@task(queue=get_task_queuename)
+def migrate_jsonfield(table, pkfield, columns):
+    batchsize = 10000
+    with advisory_lock(f'json_migration_{table}', wait=False) as acquired:
+        if not acquired:
+            return
+
+        from django.db.migrations.executor import MigrationExecutor
+
+        # If Django is currently running migrations, wait until it is done.
+        while True:
+            executor = MigrationExecutor(connection)
+            if not executor.migration_plan(executor.loader.graph.leaf_nodes()):
+                break
+            time.sleep(120)
+
+        logger.warning(f"Migrating json fields for {table}: {', '.join(columns)}")
+
+        with connection.cursor() as cursor:
+            for i in itertools.count(0, batchsize):
+                # Are there even any rows in the table beyond this point?
+                cursor.execute(f"select count(1) from {table} where {pkfield} >= %s limit 1;", (i,))
+                if not cursor.fetchone()[0]:
+                    break
+
+                column_expr = ', '.join(f"{colname} = {colname}_old::jsonb" for colname in columns)
+                # If any of the old columns have non-null values, the data needs to be cast and copied over.
+                empty_expr = ' or '.join(f"{colname}_old is not null" for colname in columns)
+                cursor.execute(  # Only clobber the new fields if there is non-null data in the old ones.
+                    f"""
+                    update {table}
+                      set {column_expr}
+                      where {pkfield} >= %s and {pkfield} < %s
+                        and {empty_expr};
+                    """,
+                    (i, i + batchsize),
+                )
+                rows = cursor.rowcount
+                logger.debug(f"Batch {i} to {i + batchsize} copied on {table}, {rows} rows affected.")
+
+            column_expr = ', '.join(f"DROP COLUMN {column}_old" for column in columns)
+            cursor.execute(f"ALTER TABLE {table} {column_expr};")
+
+        logger.warning(f"Migration of {table} to jsonb is finished.")
+
+
 @task(queue=get_task_queuename)
 def apply_cluster_membership_policies():
     from awx.main.signals import disable_activity_stream
@@ -316,13 +369,8 @@ def send_notifications(notification_list, job_id=None):
 @task(queue=get_task_queuename)
 def gather_analytics():
     from awx.conf.models import Setting
-    from rest_framework.fields import DateTimeField
 
-    last_gather = Setting.objects.filter(key='AUTOMATION_ANALYTICS_LAST_GATHER').first()
-    last_time = DateTimeField().to_internal_value(last_gather.value) if last_gather and last_gather.value else None
-    gather_time = now()
-
-    if not last_time or ((gather_time - last_time).total_seconds() > settings.AUTOMATION_ANALYTICS_GATHER_INTERVAL):
+    if is_run_threshold_reached(Setting.objects.filter(key='AUTOMATION_ANALYTICS_LAST_GATHER').first(), settings.AUTOMATION_ANALYTICS_GATHER_INTERVAL):
         analytics.gather()
 
 
@@ -381,16 +429,25 @@ def cleanup_images_and_files():
 
 @task(queue=get_task_queuename)
 def cleanup_host_metrics():
+    """Run cleanup host metrics ~each month"""
+    # TODO: move whole method to host_metrics in follow-up PR
     from awx.conf.models import Setting
+
+    if is_run_threshold_reached(
+        Setting.objects.filter(key='CLEANUP_HOST_METRICS_LAST_TS').first(), getattr(settings, 'CLEANUP_HOST_METRICS_INTERVAL', 30) * 86400
+    ):
+        months_ago = getattr(settings, 'CLEANUP_HOST_METRICS_SOFT_THRESHOLD', 12)
+        logger.info("Executing cleanup_host_metrics")
+        HostMetric.cleanup_task(months_ago)
+        logger.info("Finished cleanup_host_metrics")
+
+
+def is_run_threshold_reached(setting, threshold_seconds):
     from rest_framework.fields import DateTimeField
 
-    last_cleanup = Setting.objects.filter(key='CLEANUP_HOST_METRICS_LAST_TS').first()
-    last_time = DateTimeField().to_internal_value(last_cleanup.value) if last_cleanup and last_cleanup.value else None
+    last_time = DateTimeField().to_internal_value(setting.value) if setting and setting.value else DateTimeField().to_internal_value('1970-01-01')
 
-    cleanup_interval_secs = getattr(settings, 'CLEANUP_HOST_METRICS_INTERVAL', 30) * 86400
-    if not last_time or ((now() - last_time).total_seconds() > cleanup_interval_secs):
-        months_ago = getattr(settings, 'CLEANUP_HOST_METRICS_THRESHOLD', 12)
-        HostMetric.cleanup_task(months_ago)
+    return (now() - last_time).total_seconds() > threshold_seconds
 
 
 @task(queue=get_task_queuename)
@@ -541,7 +598,7 @@ def cluster_node_heartbeat(dispatch_time=None, worker_tasks=None):
             logger.warning(f'Heartbeat skew - interval={(nowtime - last_last_seen).total_seconds():.4f}, expected={settings.CLUSTER_NODE_HEARTBEAT_PERIOD}')
     else:
         if settings.AWX_AUTO_DEPROVISION_INSTANCES:
-            (changed, this_inst) = Instance.objects.register(ip_address=os.environ.get('MY_POD_IP'), node_type='control', uuid=settings.SYSTEM_UUID)
+            (changed, this_inst) = Instance.objects.register(ip_address=os.environ.get('MY_POD_IP'), node_type='control', node_uuid=settings.SYSTEM_UUID)
             if changed:
                 logger.warning(f'Recreated instance record {this_inst.hostname} after unexpected removal')
             this_inst.local_health_check()
@@ -839,10 +896,7 @@ def delete_inventory(inventory_id, user_id, retries=5):
             user = None
     with ignore_inventory_computed_fields(), ignore_inventory_group_removal(), impersonate(user):
         try:
-            i = Inventory.objects.get(id=inventory_id)
-            for host in i.hosts.iterator():
-                host.job_events_as_primary_host.update(host=None)
-            i.delete()
+            Inventory.objects.get(id=inventory_id).delete()
             emit_channel_notification('inventories-status_changed', {'group_name': 'inventories', 'inventory_id': inventory_id, 'status': 'deleted'})
             logger.debug('Deleted inventory {} as user {}.'.format(inventory_id, user_id))
         except Inventory.DoesNotExist:

awx/main/tests/docs/test_swagger_generation.py

@@ -7,7 +7,7 @@ from django.core.serializers.json import DjangoJSONEncoder
 from django.utils.functional import Promise
 from django.utils.encoding import force_str
 
-from openapi_codec.encode import generate_swagger_object
+from drf_yasg.codecs import OpenAPICodecJson
 import pytest
 
 from awx.api.versioning import drf_reverse
@@ -43,12 +43,12 @@ class TestSwaggerGeneration:
     @pytest.fixture(autouse=True, scope='function')
     def _prepare(self, get, admin):
         if not self.__class__.JSON:
-            url = drf_reverse('api:swagger_view') + '?format=openapi'
+            url = drf_reverse('api:schema-swagger-ui') + '?format=openapi'
             response = get(url, user=admin)
-            data = generate_swagger_object(response.data)
+            codec = OpenAPICodecJson([])
+            data = codec.generate_swagger_object(response.data)
             if response.has_header('X-Deprecated-Paths'):
                 data['deprecated_paths'] = json.loads(response['X-Deprecated-Paths'])
-            data.update(response.accepted_renderer.get_customizations() or {})
 
             data['host'] = None
             data['schemes'] = ['https']
@@ -60,12 +60,21 @@ class TestSwaggerGeneration:
                 # change {version} in paths to the actual default API version (e.g., v2)
                 revised_paths[path.replace('{version}', settings.REST_FRAMEWORK['DEFAULT_VERSION'])] = node
                 for method in node:
+                    # Ignore any parameters methods, these cause issues because it can come as an array instead of a dict
+                    # Which causes issues in the last for loop in here
+                    if method == 'parameters':
+                        continue
+
                     if path in deprecated_paths:
                         node[method]['deprecated'] = True
                     if 'description' in node[method]:
                         # Pop off the first line and use that as the summary
                         lines = node[method]['description'].splitlines()
-                        node[method]['summary'] = lines.pop(0).strip('#:')
+                        # If there was a description then set the summary as the description, otherwise make something up
+                        if lines:
+                            node[method]['summary'] = lines.pop(0).strip('#:')
+                        else:
+                            node[method]['summary'] = f'No Description for {method} on {path}'
                         node[method]['description'] = '\n'.join(lines)
 
                     # remove the required `version` parameter
@@ -90,13 +99,13 @@ class TestSwaggerGeneration:
         # The number of API endpoints changes over time, but let's just check
         # for a reasonable number here; if this test starts failing, raise/lower the bounds
         paths = JSON['paths']
-        assert 250 < len(paths) < 350
-        assert list(paths['/api/'].keys()) == ['get']
-        assert list(paths['/api/v2/'].keys()) == ['get']
-        assert list(sorted(paths['/api/v2/credentials/'].keys())) == ['get', 'post']
-        assert list(sorted(paths['/api/v2/credentials/{id}/'].keys())) == ['delete', 'get', 'patch', 'put']
-        assert list(paths['/api/v2/settings/'].keys()) == ['get']
-        assert list(paths['/api/v2/settings/{category_slug}/'].keys()) == ['get', 'put', 'patch', 'delete']
+        assert 250 < len(paths) < 375
+        assert set(list(paths['/api/'].keys())) == set(['get', 'parameters'])
+        assert set(list(paths['/api/v2/'].keys())) == set(['get', 'parameters'])
+        assert set(list(sorted(paths['/api/v2/credentials/'].keys()))) == set(['get', 'post', 'parameters'])
+        assert set(list(sorted(paths['/api/v2/credentials/{id}/'].keys()))) == set(['delete', 'get', 'patch', 'put', 'parameters'])
+        assert set(list(paths['/api/v2/settings/'].keys())) == set(['get', 'parameters'])
+        assert set(list(paths['/api/v2/settings/{category_slug}/'].keys())) == set(['get', 'put', 'patch', 'delete', 'parameters'])
 
     @pytest.mark.parametrize(
         'path',
@@ -162,4 +171,8 @@ class TestSwaggerGeneration:
             data = re.sub(r'[0-9]{4}-[0-9]{2}-[0-9]{2}(T|\s)[0-9]{2}:[0-9]{2}:[0-9]{2}.[0-9]+(Z|\+[0-9]{2}:[0-9]{2})?', r'2018-02-01T08:00:00.000000Z', data)
             data = re.sub(r'''(\s+"client_id": ")([a-zA-Z0-9]{40})("\,\s*)''', r'\1xxxx\3', data)
             data = re.sub(r'"action_node": "[^"]+"', '"action_node": "awx"', data)
+
+            # replace uuids to prevent needless diffs
+            pattern = r'[0-9a-f]{8}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{4}-[0-9a-f]{12}'
+            data = re.sub(pattern, r'00000000-0000-0000-0000-000000000000', data)
             f.write(data)

awx/main/tests/factories/fixtures.py

@@ -1,6 +1,9 @@
 import json
 
 from django.contrib.auth.models import User
+from django.core.exceptions import ValidationError
+
+from unittest import mock
 
 from awx.main.models import (
     Organization,
@@ -20,6 +23,7 @@ from awx.main.models import (
     WorkflowJobNode,
     WorkflowJobTemplateNode,
 )
+from awx.main.models.inventory import HostMetric, HostMetricSummaryMonthly
 
 # mk methods should create only a single object of a single type.
 # they should also have the option of being persisted or not.
@@ -248,3 +252,42 @@ def mk_workflow_job_node(unified_job_template=None, success_nodes=None, failure_
     if persisted:
         workflow_node.save()
     return workflow_node
+
+
+def mk_host_metric(hostname, first_automation, last_automation=None, last_deleted=None, deleted=False, persisted=True):
+    ok, idx = False, 1
+    while not ok:
+        try:
+            with mock.patch("django.utils.timezone.now") as mock_now:
+                mock_now.return_value = first_automation
+                metric = HostMetric(
+                    hostname=hostname or f"host-{first_automation}-{idx}",
+                    first_automation=first_automation,
+                    last_automation=last_automation or first_automation,
+                    last_deleted=last_deleted,
+                    deleted=deleted,
+                )
+                metric.validate_unique()
+                if persisted:
+                    metric.save()
+                ok = True
+        except ValidationError as e:
+            # Repeat create for auto-generated hostname
+            if not hostname and e.message_dict.get('hostname', None):
+                idx += 1
+            else:
+                raise e
+
+
+def mk_host_metric_summary(date, license_consumed=0, license_capacity=0, hosts_added=0, hosts_deleted=0, indirectly_managed_hosts=0, persisted=True):
+    summary = HostMetricSummaryMonthly(
+        date=date,
+        license_consumed=license_consumed,
+        license_capacity=license_capacity,
+        hosts_added=hosts_added,
+        hosts_deleted=hosts_deleted,
+        indirectly_managed_hosts=indirectly_managed_hosts,
+    )
+    if persisted:
+        summary.save()
+    return summary

awx/main/tests/functional/analytics/test_collectors.py

@@ -2,8 +2,8 @@ import pytest
 import tempfile
 import os
 import re
-import shutil
 import csv
+from io import StringIO
 
 from django.utils.timezone import now
 from datetime import timedelta
@@ -20,15 +20,16 @@ from awx.main.models import (
 )
 
 
-@pytest.fixture
-def sqlite_copy_expert(request):
-    # copy_expert is postgres-specific, and SQLite doesn't support it; mock its
-    # behavior to test that it writes a file that contains stdout from events
-    path = tempfile.mkdtemp(prefix="copied_tables")
+class MockCopy:
+    headers = None
+    results = None
+    sent_data = False
 
-    def write_stdout(self, sql, fd):
+    def __init__(self, sql, parent_connection):
         # Would be cool if we instead properly disected the SQL query and verified
         # it that way. But instead, we just take the naive approach here.
+        self.results = None
+        self.headers = None
         sql = sql.strip()
         assert sql.startswith("COPY (")
         assert sql.endswith(") TO STDOUT WITH CSV HEADER")
@@ -51,29 +52,49 @@ def sqlite_copy_expert(request):
             elif not line.endswith(","):
                 sql_new[-1] = sql_new[-1].rstrip(",")
         sql = "\n".join(sql_new)
+        parent_connection.execute(sql)
+        self.results = parent_connection.fetchall()
+        self.headers = [i[0] for i in parent_connection.description]
 
-        self.execute(sql)
-        results = self.fetchall()
-        headers = [i[0] for i in self.description]
+    def read(self):
+        if not self.sent_data:
+            mem_file = StringIO()
+            csv_handle = csv.writer(
+                mem_file,
+                delimiter=",",
+                quoting=csv.QUOTE_ALL,
+                escapechar="\\",
+                lineterminator="\n",
+            )
+            if self.headers:
+                csv_handle.writerow(self.headers)
+            if self.results:
+                csv_handle.writerows(self.results)
+            self.sent_data = True
+            return memoryview((mem_file.getvalue()).encode())
+        return None
 
-        csv_handle = csv.writer(
-            fd,
-            delimiter=",",
-            quoting=csv.QUOTE_ALL,
-            escapechar="\\",
-            lineterminator="\n",
-        )
-        csv_handle.writerow(headers)
-        csv_handle.writerows(results)
+    def __enter__(self):
+        return self
 
-    setattr(SQLiteCursorWrapper, "copy_expert", write_stdout)
-    request.addfinalizer(lambda: shutil.rmtree(path))
-    request.addfinalizer(lambda: delattr(SQLiteCursorWrapper, "copy_expert"))
-    return path
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        pass
+
+
+@pytest.fixture
+def sqlite_copy(request, mocker):
+    # copy is postgres-specific, and SQLite doesn't support it; mock its
+    # behavior to test that it writes a file that contains stdout from events
+
+    def write_stdout(self, sql):
+        mock_copy = MockCopy(sql, self)
+        return mock_copy
+
+    mocker.patch.object(SQLiteCursorWrapper, 'copy', write_stdout, create=True)
 
 
 @pytest.mark.django_db
-def test_copy_tables_unified_job_query(sqlite_copy_expert, project, inventory, job_template):
+def test_copy_tables_unified_job_query(sqlite_copy, project, inventory, job_template):
     """
     Ensure that various unified job types are in the output of the query.
     """
@@ -127,7 +148,7 @@ def workflow_job(states=["new", "new", "new", "new", "new"]):
 
 
 @pytest.mark.django_db
-def test_copy_tables_workflow_job_node_query(sqlite_copy_expert, workflow_job):
+def test_copy_tables_workflow_job_node_query(sqlite_copy, workflow_job):
     time_start = now() - timedelta(hours=9)
 
     with tempfile.TemporaryDirectory() as tmpdir:

awx/main/tests/functional/api/test_job.py

@@ -224,7 +224,7 @@ class TestControllerNode:
         return AdHocCommand.objects.create(inventory=inventory)
 
     @pytest.mark.django_db
-    def test_field_controller_node_exists(self, sqlite_copy_expert, admin_user, job, project_update, inventory_update, adhoc, get, system_job_factory):
+    def test_field_controller_node_exists(self, sqlite_copy, admin_user, job, project_update, inventory_update, adhoc, get, system_job_factory):
         system_job = system_job_factory()
 
         r = get(reverse('api:unified_job_list') + '?id={}'.format(job.id), admin_user, expect=200)

awx/main/tests/functional/api/test_unified_jobs_stdout.py

@@ -57,7 +57,7 @@ def _mk_inventory_update(created=None):
         [_mk_inventory_update, InventoryUpdateEvent, 'inventory_update', 'api:inventory_update_stdout'],
     ],
 )
-def test_text_stdout(sqlite_copy_expert, Parent, Child, relation, view, get, admin):
+def test_text_stdout(sqlite_copy, Parent, Child, relation, view, get, admin):
     job = Parent()
     job.save()
     for i in range(3):
@@ -79,7 +79,7 @@ def test_text_stdout(sqlite_copy_expert, Parent, Child, relation, view, get, adm
     ],
 )
 @pytest.mark.parametrize('download', [True, False])
-def test_ansi_stdout_filtering(sqlite_copy_expert, Parent, Child, relation, view, download, get, admin):
+def test_ansi_stdout_filtering(sqlite_copy, Parent, Child, relation, view, download, get, admin):
     job = Parent()
     job.save()
     for i in range(3):
@@ -111,7 +111,7 @@ def test_ansi_stdout_filtering(sqlite_copy_expert, Parent, Child, relation, view
         [_mk_inventory_update, InventoryUpdateEvent, 'inventory_update', 'api:inventory_update_stdout'],
     ],
 )
-def test_colorized_html_stdout(sqlite_copy_expert, Parent, Child, relation, view, get, admin):
+def test_colorized_html_stdout(sqlite_copy, Parent, Child, relation, view, get, admin):
     job = Parent()
     job.save()
     for i in range(3):
@@ -134,7 +134,7 @@ def test_colorized_html_stdout(sqlite_copy_expert, Parent, Child, relation, view
         [_mk_inventory_update, InventoryUpdateEvent, 'inventory_update', 'api:inventory_update_stdout'],
     ],
 )
-def test_stdout_line_range(sqlite_copy_expert, Parent, Child, relation, view, get, admin):
+def test_stdout_line_range(sqlite_copy, Parent, Child, relation, view, get, admin):
     job = Parent()
     job.save()
     for i in range(20):
@@ -146,7 +146,7 @@ def test_stdout_line_range(sqlite_copy_expert, Parent, Child, relation, view, ge
 
 
 @pytest.mark.django_db
-def test_text_stdout_from_system_job_events(sqlite_copy_expert, get, admin):
+def test_text_stdout_from_system_job_events(sqlite_copy, get, admin):
     created = tz_now()
     job = SystemJob(created=created)
     job.save()
@@ -158,7 +158,7 @@ def test_text_stdout_from_system_job_events(sqlite_copy_expert, get, admin):
 
 
 @pytest.mark.django_db
-def test_text_stdout_with_max_stdout(sqlite_copy_expert, get, admin):
+def test_text_stdout_with_max_stdout(sqlite_copy, get, admin):
     created = tz_now()
     job = SystemJob(created=created)
     job.save()
@@ -185,7 +185,7 @@ def test_text_stdout_with_max_stdout(sqlite_copy_expert, get, admin):
 )
 @pytest.mark.parametrize('fmt', ['txt', 'ansi'])
 @mock.patch('awx.main.redact.UriCleaner.SENSITIVE_URI_PATTERN', mock.Mock(**{'search.return_value': None}))  # really slow for large strings
-def test_max_bytes_display(sqlite_copy_expert, Parent, Child, relation, view, fmt, get, admin):
+def test_max_bytes_display(sqlite_copy, Parent, Child, relation, view, fmt, get, admin):
     created = tz_now()
     job = Parent(created=created)
     job.save()
@@ -255,7 +255,7 @@ def test_legacy_result_stdout_with_max_bytes(Cls, view, fmt, get, admin):
     ],
 )
 @pytest.mark.parametrize('fmt', ['txt', 'ansi', 'txt_download', 'ansi_download'])
-def test_text_with_unicode_stdout(sqlite_copy_expert, Parent, Child, relation, view, get, admin, fmt):
+def test_text_with_unicode_stdout(sqlite_copy, Parent, Child, relation, view, get, admin, fmt):
     job = Parent()
     job.save()
     for i in range(3):
@@ -267,7 +267,7 @@ def test_text_with_unicode_stdout(sqlite_copy_expert, Parent, Child, relation, v
 
 
 @pytest.mark.django_db
-def test_unicode_with_base64_ansi(sqlite_copy_expert, get, admin):
+def test_unicode_with_base64_ansi(sqlite_copy, get, admin):
     created = tz_now()
     job = Job(created=created)
     job.save()

awx/main/tests/functional/commands/test_host_metric_summary_monthly.py

@@ -0,0 +1,382 @@
+import pytest
+import datetime
+from dateutil.relativedelta import relativedelta
+from django.conf import settings
+from django.utils import timezone
+
+
+from awx.main.management.commands.host_metric_summary_monthly import Command
+from awx.main.models.inventory import HostMetric, HostMetricSummaryMonthly
+from awx.main.tests.factories.fixtures import mk_host_metric, mk_host_metric_summary
+
+
+@pytest.fixture
+def threshold():
+    return int(getattr(settings, 'CLEANUP_HOST_METRICS_HARD_THRESHOLD', 36))
+
+
+@pytest.mark.django_db
+@pytest.mark.parametrize("metrics_cnt", [0, 1, 2, 3])
+@pytest.mark.parametrize("mode", ["old_data", "actual_data", "all_data"])
+def test_summaries_counts(threshold, metrics_cnt, mode):
+    assert HostMetricSummaryMonthly.objects.count() == 0
+
+    for idx in range(metrics_cnt):
+        if mode == "old_data" or mode == "all_data":
+            mk_host_metric(None, months_ago(threshold + idx, "dt"))
+        elif mode == "actual_data" or mode == "all_data":
+            mk_host_metric(None, (months_ago(threshold - idx, "dt")))
+
+    Command().handle()
+
+    # Number of records is equal to host metrics' hard cleanup months
+    assert HostMetricSummaryMonthly.objects.count() == threshold
+
+    # Records start with date in the month following to the threshold month
+    date = months_ago(threshold - 1)
+    for metric in list(HostMetricSummaryMonthly.objects.order_by('date').all()):
+        assert metric.date == date
+        date += relativedelta(months=1)
+
+    # Older record are untouched
+    mk_host_metric_summary(date=months_ago(threshold + 10))
+    Command().handle()
+
+    assert HostMetricSummaryMonthly.objects.count() == threshold + 1
+
+
+@pytest.mark.django_db
+@pytest.mark.parametrize("mode", ["old_data", "actual_data", "all_data"])
+def test_summary_values(threshold, mode):
+    tester = {"old_data": MetricsTesterOldData(threshold), "actual_data": MetricsTesterActualData(threshold), "all_data": MetricsTesterCombinedData(threshold)}[
+        mode
+    ]
+
+    for iteration in ["create_metrics", "add_old_summaries", "change_metrics", "delete_metrics", "add_metrics"]:
+        getattr(tester, iteration)()  # call method by string
+
+        # Operation is idempotent, repeat twice
+        for _ in range(2):
+            Command().handle()
+            # call assert method by string
+            getattr(tester, f"assert_{iteration}")()
+
+
+class MetricsTester:
+    def __init__(self, threshold, ignore_asserts=False):
+        self.threshold = threshold
+        self.expected_summaries = {}
+        self.ignore_asserts = ignore_asserts
+
+    def add_old_summaries(self):
+        """These records don't correspond with Host metrics"""
+        mk_host_metric_summary(self.below(4), license_consumed=100, hosts_added=10, hosts_deleted=5)
+        mk_host_metric_summary(self.below(3), license_consumed=105, hosts_added=20, hosts_deleted=10)
+        mk_host_metric_summary(self.below(2), license_consumed=115, hosts_added=60, hosts_deleted=75)
+
+    def assert_add_old_summaries(self):
+        """Old summary records should be untouched"""
+        self.expected_summaries[self.below(4)] = {"date": self.below(4), "license_consumed": 100, "hosts_added": 10, "hosts_deleted": 5}
+        self.expected_summaries[self.below(3)] = {"date": self.below(3), "license_consumed": 105, "hosts_added": 20, "hosts_deleted": 10}
+        self.expected_summaries[self.below(2)] = {"date": self.below(2), "license_consumed": 115, "hosts_added": 60, "hosts_deleted": 75}
+
+        self.assert_host_metric_summaries()
+
+    def assert_host_metric_summaries(self):
+        """Ignore asserts when old/actual test object is used only as a helper for Combined test"""
+        if self.ignore_asserts:
+            return True
+
+        for summary in list(HostMetricSummaryMonthly.objects.order_by('date').all()):
+            assert self.expected_summaries.get(summary.date, None) is not None
+
+            assert self.expected_summaries[summary.date] == {
+                "date": summary.date,
+                "license_consumed": summary.license_consumed,
+                "hosts_added": summary.hosts_added,
+                "hosts_deleted": summary.hosts_deleted,
+            }
+
+    def below(self, months, fmt="date"):
+        """months below threshold, returns first date of that month"""
+        date = months_ago(self.threshold + months)
+        if fmt == "dt":
+            return timezone.make_aware(datetime.datetime.combine(date, datetime.datetime.min.time()))
+        else:
+            return date
+
+    def above(self, months, fmt="date"):
+        """months above threshold, returns first date of that month"""
+        date = months_ago(self.threshold - months)
+        if fmt == "dt":
+            return timezone.make_aware(datetime.datetime.combine(date, datetime.datetime.min.time()))
+        else:
+            return date
+
+
+class MetricsTesterOldData(MetricsTester):
+    def create_metrics(self):
+        """Creates 7 host metrics older than delete threshold"""
+        mk_host_metric("host_1", first_automation=self.below(3, "dt"))
+        mk_host_metric("host_2", first_automation=self.below(2, "dt"))
+        mk_host_metric("host_3", first_automation=self.below(2, "dt"), last_deleted=self.above(2, "dt"), deleted=False)
+        mk_host_metric("host_4", first_automation=self.below(2, "dt"), last_deleted=self.above(2, "dt"), deleted=True)
+        mk_host_metric("host_5", first_automation=self.below(2, "dt"), last_deleted=self.below(2, "dt"), deleted=True)
+        mk_host_metric("host_6", first_automation=self.below(1, "dt"), last_deleted=self.below(1, "dt"), deleted=False)
+        mk_host_metric("host_7", first_automation=self.below(1, "dt"))
+
+    def assert_create_metrics(self):
+        """
+        Month 1 is computed from older host metrics,
+        Month 2 has deletion (host_4)
+        Other months are unchanged (same as month 2)
+        """
+        self.expected_summaries = {
+            self.above(1): {"date": self.above(1), "license_consumed": 6, "hosts_added": 0, "hosts_deleted": 0},
+            self.above(2): {"date": self.above(2), "license_consumed": 5, "hosts_added": 0, "hosts_deleted": 1},
+        }
+        # no change in months 3+
+        idx = 3
+        month = self.above(idx)
+        while month <= beginning_of_the_month():
+            self.expected_summaries[self.above(idx)] = {"date": self.above(idx), "license_consumed": 5, "hosts_added": 0, "hosts_deleted": 0}
+            month += relativedelta(months=1)
+            idx += 1
+
+        self.assert_host_metric_summaries()
+
+    def add_old_summaries(self):
+        super().add_old_summaries()
+
+    def assert_add_old_summaries(self):
+        super().assert_add_old_summaries()
+
+    @staticmethod
+    def change_metrics():
+        """Hosts 1,2 soft deleted, host_4 automated again (undeleted)"""
+        HostMetric.objects.filter(hostname='host_1').update(last_deleted=beginning_of_the_month("dt"), deleted=True)
+        HostMetric.objects.filter(hostname='host_2').update(last_deleted=timezone.now(), deleted=True)
+        HostMetric.objects.filter(hostname='host_4').update(deleted=False)
+
+    def assert_change_metrics(self):
+        """
+        Summaries since month 2 were changed (host_4 restored == automated again)
+        Current month has 2 deletions (host_1, host_2)
+        """
+        self.expected_summaries[self.above(2)] |= {'hosts_deleted': 0}
+        for idx in range(2, self.threshold):
+            self.expected_summaries[self.above(idx)] |= {'license_consumed': 6}
+        self.expected_summaries[beginning_of_the_month()] |= {'license_consumed': 4, 'hosts_deleted': 2}
+
+        self.assert_host_metric_summaries()
+
+    @staticmethod
+    def delete_metrics():
+        """Deletes metric deleted before the threshold"""
+        HostMetric.objects.filter(hostname='host_5').delete()
+
+    def assert_delete_metrics(self):
+        """No change"""
+        self.assert_host_metric_summaries()
+
+    @staticmethod
+    def add_metrics():
+        """Adds new metrics"""
+        mk_host_metric("host_24", first_automation=beginning_of_the_month("dt"))
+        mk_host_metric("host_25", first_automation=beginning_of_the_month("dt"))  # timezone.now())
+
+    def assert_add_metrics(self):
+        """Summary in current month is updated"""
+        self.expected_summaries[beginning_of_the_month()]['license_consumed'] = 6
+        self.expected_summaries[beginning_of_the_month()]['hosts_added'] = 2
+
+        self.assert_host_metric_summaries()
+
+
+class MetricsTesterActualData(MetricsTester):
+    def create_metrics(self):
+        """Creates 16 host metrics newer than delete threshold"""
+        mk_host_metric("host_8", first_automation=self.above(1, "dt"))
+        mk_host_metric("host_9", first_automation=self.above(1, "dt"), last_deleted=self.above(1, "dt"))
+        mk_host_metric("host_10", first_automation=self.above(1, "dt"), last_deleted=self.above(1, "dt"), deleted=True)
+        mk_host_metric("host_11", first_automation=self.above(1, "dt"), last_deleted=self.above(2, "dt"))
+        mk_host_metric("host_12", first_automation=self.above(1, "dt"), last_deleted=self.above(2, "dt"), deleted=True)
+        mk_host_metric("host_13", first_automation=self.above(2, "dt"))
+        mk_host_metric("host_14", first_automation=self.above(2, "dt"), last_deleted=self.above(2, "dt"))
+        mk_host_metric("host_15", first_automation=self.above(2, "dt"), last_deleted=self.above(2, "dt"), deleted=True)
+        mk_host_metric("host_16", first_automation=self.above(2, "dt"), last_deleted=self.above(3, "dt"))
+        mk_host_metric("host_17", first_automation=self.above(2, "dt"), last_deleted=self.above(3, "dt"), deleted=True)
+        mk_host_metric("host_18", first_automation=self.above(4, "dt"))
+        # next one shouldn't happen in real (deleted=True, last_deleted = NULL)
+        mk_host_metric("host_19", first_automation=self.above(4, "dt"), deleted=True)
+        mk_host_metric("host_20", first_automation=self.above(4, "dt"), last_deleted=self.above(4, "dt"))
+        mk_host_metric("host_21", first_automation=self.above(4, "dt"), last_deleted=self.above(4, "dt"), deleted=True)
+        mk_host_metric("host_22", first_automation=self.above(4, "dt"), last_deleted=self.above(5, "dt"))
+        mk_host_metric("host_23", first_automation=self.above(4, "dt"), last_deleted=self.above(5, "dt"), deleted=True)
+
+    def assert_create_metrics(self):
+        self.expected_summaries = {
+            self.above(1): {"date": self.above(1), "license_consumed": 4, "hosts_added": 5, "hosts_deleted": 1},
+            self.above(2): {"date": self.above(2), "license_consumed": 7, "hosts_added": 5, "hosts_deleted": 2},
+            self.above(3): {"date": self.above(3), "license_consumed": 6, "hosts_added": 0, "hosts_deleted": 1},
+            self.above(4): {"date": self.above(4), "license_consumed": 11, "hosts_added": 6, "hosts_deleted": 1},
+            self.above(5): {"date": self.above(5), "license_consumed": 10, "hosts_added": 0, "hosts_deleted": 1},
+        }
+        # no change in months 6+
+        idx = 6
+        month = self.above(idx)
+        while month <= beginning_of_the_month():
+            self.expected_summaries[self.above(idx)] = {"date": self.above(idx), "license_consumed": 10, "hosts_added": 0, "hosts_deleted": 0}
+            month += relativedelta(months=1)
+            idx += 1
+
+        self.assert_host_metric_summaries()
+
+    def add_old_summaries(self):
+        super().add_old_summaries()
+
+    def assert_add_old_summaries(self):
+        super().assert_add_old_summaries()
+
+    @staticmethod
+    def change_metrics():
+        """
+        - Hosts 12, 19, 21 were automated again (undeleted)
+        - Host 16 was soft deleted
+        - Host 17 was undeleted and soft deleted again
+        """
+        HostMetric.objects.filter(hostname='host_12').update(deleted=False)
+        HostMetric.objects.filter(hostname='host_16').update(last_deleted=timezone.now(), deleted=True)
+        HostMetric.objects.filter(hostname='host_17').update(last_deleted=beginning_of_the_month("dt"), deleted=True)
+        HostMetric.objects.filter(hostname='host_19').update(deleted=False)
+        HostMetric.objects.filter(hostname='host_21').update(deleted=False)
+
+    def assert_change_metrics(self):
+        """
+        Summaries since month 2 were changed
+        Current month has 2 deletions (host_16, host_17)
+        """
+        self.expected_summaries[self.above(2)] |= {'license_consumed': 8, 'hosts_deleted': 1}
+        self.expected_summaries[self.above(3)] |= {'license_consumed': 8, 'hosts_deleted': 0}
+        self.expected_summaries[self.above(4)] |= {'license_consumed': 14, 'hosts_deleted': 0}
+
+        # month 5 had hosts_deleted 1 => license_consumed == 14 - 1
+        for idx in range(5, self.threshold):
+            self.expected_summaries[self.above(idx)] |= {'license_consumed': 13}
+        self.expected_summaries[beginning_of_the_month()] |= {'license_consumed': 11, 'hosts_deleted': 2}
+
+        self.assert_host_metric_summaries()
+
+    def delete_metrics(self):
+        """Hard cleanup can't delete metrics newer than threshold. No change"""
+        pass
+
+    def assert_delete_metrics(self):
+        """No change"""
+        self.assert_host_metric_summaries()
+
+    @staticmethod
+    def add_metrics():
+        """Adds new metrics"""
+        mk_host_metric("host_26", first_automation=beginning_of_the_month("dt"))
+        mk_host_metric("host_27", first_automation=timezone.now())
+
+    def assert_add_metrics(self):
+        """
+        Two metrics were deleted in current month by change_metrics()
+        Two metrics are added now
+        => license_consumed is equal to the previous month (13 - 2 + 2)
+        """
+        self.expected_summaries[beginning_of_the_month()] |= {'license_consumed': 13, 'hosts_added': 2}
+
+        self.assert_host_metric_summaries()
+
+
+class MetricsTesterCombinedData(MetricsTester):
+    def __init__(self, threshold):
+        super().__init__(threshold)
+        self.old_data = MetricsTesterOldData(threshold, ignore_asserts=True)
+        self.actual_data = MetricsTesterActualData(threshold, ignore_asserts=True)
+
+    def assert_host_metric_summaries(self):
+        self._combine_expected_summaries()
+        super().assert_host_metric_summaries()
+
+    def create_metrics(self):
+        self.old_data.create_metrics()
+        self.actual_data.create_metrics()
+
+    def assert_create_metrics(self):
+        self.old_data.assert_create_metrics()
+        self.actual_data.assert_create_metrics()
+
+        self.assert_host_metric_summaries()
+
+    def add_old_summaries(self):
+        super().add_old_summaries()
+
+    def assert_add_old_summaries(self):
+        self.old_data.assert_add_old_summaries()
+        self.actual_data.assert_add_old_summaries()
+
+        self.assert_host_metric_summaries()
+
+    def change_metrics(self):
+        self.old_data.change_metrics()
+        self.actual_data.change_metrics()
+
+    def assert_change_metrics(self):
+        self.old_data.assert_change_metrics()
+        self.actual_data.assert_change_metrics()
+
+        self.assert_host_metric_summaries()
+
+    def delete_metrics(self):
+        self.old_data.delete_metrics()
+        self.actual_data.delete_metrics()
+
+    def assert_delete_metrics(self):
+        self.old_data.assert_delete_metrics()
+        self.actual_data.assert_delete_metrics()
+
+        self.assert_host_metric_summaries()
+
+    def add_metrics(self):
+        self.old_data.add_metrics()
+        self.actual_data.add_metrics()
+
+    def assert_add_metrics(self):
+        self.old_data.assert_add_metrics()
+        self.actual_data.assert_add_metrics()
+
+        self.assert_host_metric_summaries()
+
+    def _combine_expected_summaries(self):
+        """
+        Expected summaries are sum of expected values for tests with old and actual data
+        Except data older than hard delete threshold (these summaries are untouched by task => the same in all tests)
+        """
+        for date, summary in self.old_data.expected_summaries.items():
+            if date <= months_ago(self.threshold):
+                license_consumed = summary['license_consumed']
+                hosts_added = summary['hosts_added']
+                hosts_deleted = summary['hosts_deleted']
+            else:
+                license_consumed = summary['license_consumed'] + self.actual_data.expected_summaries[date]['license_consumed']
+                hosts_added = summary['hosts_added'] + self.actual_data.expected_summaries[date]['hosts_added']
+                hosts_deleted = summary['hosts_deleted'] + self.actual_data.expected_summaries[date]['hosts_deleted']
+            self.expected_summaries[date] = {'date': date, 'license_consumed': license_consumed, 'hosts_added': hosts_added, 'hosts_deleted': hosts_deleted}
+
+
+def months_ago(num, fmt="date"):
+    if num is None:
+        return None
+    return beginning_of_the_month(fmt) - relativedelta(months=num)
+
+
+def beginning_of_the_month(fmt="date"):
+    date = datetime.date.today().replace(day=1)
+    if fmt == "dt":
+        return timezone.make_aware(datetime.datetime.combine(date, datetime.datetime.min.time()))
+    else:
+        return date

awx/main/tests/functional/conftest.py

@@ -1,8 +1,6 @@
 # Python
 import pytest
 from unittest import mock
-import tempfile
-import shutil
 import urllib.parse
 from unittest.mock import PropertyMock
 
@@ -789,25 +787,43 @@ def oauth_application(admin):
     return Application.objects.create(name='test app', user=admin, client_type='confidential', authorization_grant_type='password')
 
 
-@pytest.fixture
-def sqlite_copy_expert(request):
-    # copy_expert is postgres-specific, and SQLite doesn't support it; mock its
-    # behavior to test that it writes a file that contains stdout from events
-    path = tempfile.mkdtemp(prefix='job-event-stdout')
+class MockCopy:
+    events = []
+    index = -1
 
-    def write_stdout(self, sql, fd):
-        # simulate postgres copy_expert support with ORM code
+    def __init__(self, sql):
+        self.events = []
         parts = sql.split(' ')
         tablename = parts[parts.index('from') + 1]
         for cls in (JobEvent, AdHocCommandEvent, ProjectUpdateEvent, InventoryUpdateEvent, SystemJobEvent):
             if cls._meta.db_table == tablename:
                 for event in cls.objects.order_by('start_line').all():
-                    fd.write(event.stdout)
+                    self.events.append(event.stdout)
 
-    setattr(SQLiteCursorWrapper, 'copy_expert', write_stdout)
-    request.addfinalizer(lambda: shutil.rmtree(path))
-    request.addfinalizer(lambda: delattr(SQLiteCursorWrapper, 'copy_expert'))
-    return path
+    def read(self):
+        self.index = self.index + 1
+        if self.index < len(self.events):
+            return memoryview(self.events[self.index].encode())
+
+        return None
+
+    def __enter__(self):
+        return self
+
+    def __exit__(self, exc_type, exc_val, exc_tb):
+        pass
+
+
+@pytest.fixture
+def sqlite_copy(request, mocker):
+    # copy is postgres-specific, and SQLite doesn't support it; mock its
+    # behavior to test that it writes a file that contains stdout from events
+
+    def write_stdout(self, sql):
+        mock_copy = MockCopy(sql)
+        return mock_copy
+
+    mocker.patch.object(SQLiteCursorWrapper, 'copy', write_stdout, create=True)
 
 
 @pytest.fixture

awx/main/tests/functional/models/test_notifications.py

@@ -98,7 +98,7 @@ class TestJobNotificationMixin(object):
 
     @pytest.mark.django_db
     @pytest.mark.parametrize('JobClass', [AdHocCommand, InventoryUpdate, Job, ProjectUpdate, SystemJob, WorkflowJob])
-    def test_context(self, JobClass, sqlite_copy_expert, project, inventory_source):
+    def test_context(self, JobClass, sqlite_copy, project, inventory_source):
         """The Jinja context defines all of the fields that can be used by a template. Ensure that the context generated
         for each job type has the expected structure."""
         kwargs = {}

awx/main/tests/functional/task_management/test_scheduler.py

@@ -331,15 +331,13 @@ def test_single_job_dependencies_project_launch(controlplane_instance_group, job
     p.save(skip_update=True)
     with mock.patch("awx.main.scheduler.TaskManager.start_task"):
         dm = DependencyManager()
-        with mock.patch.object(DependencyManager, "create_project_update", wraps=dm.create_project_update) as mock_pu:
-            dm.schedule()
-            mock_pu.assert_called_once_with(j)
-            pu = [x for x in p.project_updates.all()]
-            assert len(pu) == 1
-            TaskManager().schedule()
-            TaskManager.start_task.assert_called_once_with(pu[0], controlplane_instance_group, instance)
-            pu[0].status = "successful"
-            pu[0].save()
+        dm.schedule()
+        pu = [x for x in p.project_updates.all()]
+        assert len(pu) == 1
+        TaskManager().schedule()
+        TaskManager.start_task.assert_called_once_with(pu[0], controlplane_instance_group, instance)
+        pu[0].status = "successful"
+        pu[0].save()
     with mock.patch("awx.main.scheduler.TaskManager.start_task"):
         TaskManager().schedule()
         TaskManager.start_task.assert_called_once_with(j, controlplane_instance_group, instance)
@@ -359,15 +357,14 @@ def test_single_job_dependencies_inventory_update_launch(controlplane_instance_g
     i.inventory_sources.add(ii)
     with mock.patch("awx.main.scheduler.TaskManager.start_task"):
         dm = DependencyManager()
-        with mock.patch.object(DependencyManager, "create_inventory_update", wraps=dm.create_inventory_update) as mock_iu:
-            dm.schedule()
-            mock_iu.assert_called_once_with(j, ii)
-            iu = [x for x in ii.inventory_updates.all()]
-            assert len(iu) == 1
-            TaskManager().schedule()
-            TaskManager.start_task.assert_called_once_with(iu[0], controlplane_instance_group, instance)
-            iu[0].status = "successful"
-            iu[0].save()
+        dm.schedule()
+        assert ii.inventory_updates.count() == 1
+        iu = [x for x in ii.inventory_updates.all()]
+        assert len(iu) == 1
+        TaskManager().schedule()
+        TaskManager.start_task.assert_called_once_with(iu[0], controlplane_instance_group, instance)
+        iu[0].status = "successful"
+        iu[0].save()
     with mock.patch("awx.main.scheduler.TaskManager.start_task"):
         TaskManager().schedule()
         TaskManager.start_task.assert_called_once_with(j, controlplane_instance_group, instance)
@@ -382,11 +379,11 @@ def test_inventory_update_launches_project_update(controlplane_instance_group, s
     iu = ii.create_inventory_update()
     iu.status = "pending"
     iu.save()
+    assert project.project_updates.count() == 0
     with mock.patch("awx.main.scheduler.TaskManager.start_task"):
         dm = DependencyManager()
-        with mock.patch.object(DependencyManager, "create_project_update", wraps=dm.create_project_update) as mock_pu:
-            dm.schedule()
-            mock_pu.assert_called_with(iu, project_id=project.id)
+        dm.schedule()
+    assert project.project_updates.count() == 1
 
 
 @pytest.mark.django_db
@@ -407,9 +404,8 @@ def test_job_dependency_with_already_updated(controlplane_instance_group, job_te
     j.save()
     with mock.patch("awx.main.scheduler.TaskManager.start_task"):
         dm = DependencyManager()
-        with mock.patch.object(DependencyManager, "create_inventory_update", wraps=dm.create_inventory_update) as mock_iu:
-            dm.schedule()
-            mock_iu.assert_not_called()
+        dm.schedule()
+        assert ii.inventory_updates.count() == 0
     with mock.patch("awx.main.scheduler.TaskManager.start_task"):
         TaskManager().schedule()
         TaskManager.start_task.assert_called_once_with(j, controlplane_instance_group, instance)
@@ -442,7 +438,9 @@ def test_shared_dependencies_launch(controlplane_instance_group, job_template_fa
         TaskManager().schedule()
         pu = p.project_updates.first()
         iu = ii.inventory_updates.first()
-        TaskManager.start_task.assert_has_calls([mock.call(iu, controlplane_instance_group, instance), mock.call(pu, controlplane_instance_group, instance)])
+        TaskManager.start_task.assert_has_calls(
+            [mock.call(iu, controlplane_instance_group, instance), mock.call(pu, controlplane_instance_group, instance)], any_order=True
+        )
         pu.status = "successful"
         pu.finished = pu.created + timedelta(seconds=1)
         pu.save()
@@ -451,7 +449,9 @@ def test_shared_dependencies_launch(controlplane_instance_group, job_template_fa
         iu.save()
     with mock.patch("awx.main.scheduler.TaskManager.start_task"):
         TaskManager().schedule()
-        TaskManager.start_task.assert_has_calls([mock.call(j1, controlplane_instance_group, instance), mock.call(j2, controlplane_instance_group, instance)])
+        TaskManager.start_task.assert_has_calls(
+            [mock.call(j1, controlplane_instance_group, instance), mock.call(j2, controlplane_instance_group, instance)], any_order=True
+        )
     pu = [x for x in p.project_updates.all()]
     iu = [x for x in ii.inventory_updates.all()]
     assert len(pu) == 1

awx/main/tests/functional/test_credential.py

@@ -78,6 +78,7 @@ def test_default_cred_types():
         [
             'aim',
             'aws',
+            'aws_secretsmanager_credential',
             'azure_kv',
             'azure_rm',
             'centrify_vault_kv',

awx/main/tests/functional/test_dispatch.py

@@ -3,6 +3,7 @@ import multiprocessing
 import random
 import signal
 import time
+import yaml
 from unittest import mock
 
 from django.utils.timezone import now as tz_now
@@ -13,6 +14,7 @@ from awx.main.dispatch import reaper
 from awx.main.dispatch.pool import StatefulPoolWorker, WorkerPool, AutoscalePool
 from awx.main.dispatch.publish import task
 from awx.main.dispatch.worker import BaseWorker, TaskWorker
+from awx.main.dispatch.periodic import Scheduler
 
 
 '''
@@ -439,3 +441,76 @@ class TestJobReaper(object):
         assert job.started > ref_time
         assert job.status == 'running'
         assert job.job_explanation == ''
+
+
+@pytest.mark.django_db
+class TestScheduler:
+    def test_too_many_schedules_freak_out(self):
+        with pytest.raises(RuntimeError):
+            Scheduler({'job1': {'schedule': datetime.timedelta(seconds=1)}, 'job2': {'schedule': datetime.timedelta(seconds=1)}})
+
+    def test_spread_out(self):
+        scheduler = Scheduler(
+            {
+                'job1': {'schedule': datetime.timedelta(seconds=16)},
+                'job2': {'schedule': datetime.timedelta(seconds=16)},
+                'job3': {'schedule': datetime.timedelta(seconds=16)},
+                'job4': {'schedule': datetime.timedelta(seconds=16)},
+            }
+        )
+        assert [job.offset for job in scheduler.jobs] == [0, 4, 8, 12]
+
+    def test_missed_schedule(self, mocker):
+        scheduler = Scheduler({'job1': {'schedule': datetime.timedelta(seconds=10)}})
+        assert scheduler.jobs[0].missed_runs(time.time() - scheduler.global_start) == 0
+        mocker.patch('awx.main.dispatch.periodic.time.time', return_value=scheduler.global_start + 50)
+        scheduler.get_and_mark_pending()
+        assert scheduler.jobs[0].missed_runs(50) > 1
+
+    def test_advance_schedule(self, mocker):
+        scheduler = Scheduler(
+            {
+                'job1': {'schedule': datetime.timedelta(seconds=30)},
+                'joba': {'schedule': datetime.timedelta(seconds=20)},
+                'jobb': {'schedule': datetime.timedelta(seconds=20)},
+            }
+        )
+        for job in scheduler.jobs:
+            # HACK: the offsets automatically added make this a hard test to write... so remove offsets
+            job.offset = 0.0
+        mocker.patch('awx.main.dispatch.periodic.time.time', return_value=scheduler.global_start + 29)
+        to_run = scheduler.get_and_mark_pending()
+        assert set(job.name for job in to_run) == set(['joba', 'jobb'])
+        mocker.patch('awx.main.dispatch.periodic.time.time', return_value=scheduler.global_start + 39)
+        to_run = scheduler.get_and_mark_pending()
+        assert len(to_run) == 1
+        assert to_run[0].name == 'job1'
+
+    @staticmethod
+    def get_job(scheduler, name):
+        for job in scheduler.jobs:
+            if job.name == name:
+                return job
+
+    def test_scheduler_debug(self, mocker):
+        scheduler = Scheduler(
+            {
+                'joba': {'schedule': datetime.timedelta(seconds=20)},
+                'jobb': {'schedule': datetime.timedelta(seconds=50)},
+                'jobc': {'schedule': datetime.timedelta(seconds=500)},
+                'jobd': {'schedule': datetime.timedelta(seconds=20)},
+            }
+        )
+        rel_time = 119.9  # slightly under the 6th 20-second bin, to avoid offset problems
+        current_time = scheduler.global_start + rel_time
+        mocker.patch('awx.main.dispatch.periodic.time.time', return_value=current_time - 1.0e-8)
+        self.get_job(scheduler, 'jobb').mark_run(rel_time)
+        self.get_job(scheduler, 'jobd').mark_run(rel_time - 20.0)
+
+        output = scheduler.debug()
+        data = yaml.safe_load(output)
+        assert data['schedule_list']['jobc']['last_run_seconds_ago'] is None
+        assert data['schedule_list']['joba']['missed_runs'] == 4
+        assert data['schedule_list']['jobd']['missed_runs'] == 3
+        assert data['schedule_list']['jobd']['completed_runs'] == 1
+        assert data['schedule_list']['jobb']['next_run_in_seconds'] > 25.0

awx/main/tests/functional/test_jobs.py

@@ -6,6 +6,7 @@ import json
 from awx.main.models import (
     Job,
     Instance,
+    Host,
     JobHostSummary,
     InventoryUpdate,
     InventorySource,
@@ -18,6 +19,9 @@ from awx.main.models import (
     ExecutionEnvironment,
 )
 from awx.main.tasks.system import cluster_node_heartbeat
+from awx.main.tasks.facts import update_hosts
+
+from django.db import OperationalError
 from django.test.utils import override_settings
 
 
@@ -112,6 +116,51 @@ def test_job_notification_host_data(inventory, machine_credential, project, job_
     }
 
 
+@pytest.mark.django_db
+class TestAnsibleFactsSave:
+    current_call = 0
+
+    def test_update_hosts_deleted_host(self, inventory):
+        hosts = [Host.objects.create(inventory=inventory, name=f'foo{i}') for i in range(3)]
+        for host in hosts:
+            host.ansible_facts = {'foo': 'bar'}
+        last_pk = hosts[-1].pk
+        assert inventory.hosts.count() == 3
+        Host.objects.get(pk=last_pk).delete()
+        assert inventory.hosts.count() == 2
+        update_hosts(hosts)
+        assert inventory.hosts.count() == 2
+        for host in inventory.hosts.all():
+            host.refresh_from_db()
+            assert host.ansible_facts == {'foo': 'bar'}
+
+    def test_update_hosts_forever_deadlock(self, inventory, mocker):
+        hosts = [Host.objects.create(inventory=inventory, name=f'foo{i}') for i in range(3)]
+        for host in hosts:
+            host.ansible_facts = {'foo': 'bar'}
+        db_mock = mocker.patch('awx.main.tasks.facts.Host.objects.bulk_update')
+        db_mock.side_effect = OperationalError('deadlock detected')
+        with pytest.raises(OperationalError):
+            update_hosts(hosts)
+
+    def fake_bulk_update(self, host_list):
+        if self.current_call > 2:
+            return Host.objects.bulk_update(host_list, ['ansible_facts', 'ansible_facts_modified'])
+        self.current_call += 1
+        raise OperationalError('deadlock detected')
+
+    def test_update_hosts_resolved_deadlock(self, inventory, mocker):
+        hosts = [Host.objects.create(inventory=inventory, name=f'foo{i}') for i in range(3)]
+        for host in hosts:
+            host.ansible_facts = {'foo': 'bar'}
+        self.current_call = 0
+        mocker.patch('awx.main.tasks.facts.raw_update_hosts', new=self.fake_bulk_update)
+        update_hosts(hosts)
+        for host in inventory.hosts.all():
+            host.refresh_from_db()
+            assert host.ansible_facts == {'foo': 'bar'}
+
+
 @pytest.mark.django_db
 class TestLaunchConfig:
     def test_null_creation_from_prompts(self):

awx/main/tests/functional/test_notifications.py

@@ -16,8 +16,7 @@ from django.test.utils import override_settings
 @pytest.mark.django_db
 def test_get_notification_template_list(get, user, notification_template):
     url = reverse('api:notification_template_list')
-    response = get(url, user('admin', True))
-    assert response.status_code == 200
+    response = get(url, user('admin', True), expect=200)
     assert len(response.data['results']) == 1
 
 
@@ -35,8 +34,8 @@ def test_basic_parameterization(get, post, user, organization):
             notification_configuration=dict(url="http://localhost", disable_ssl_verification=False, headers={"Test": "Header"}),
         ),
         u,
+        expect=201,
     )
-    assert response.status_code == 201
     url = reverse('api:notification_template_detail', kwargs={'pk': response.data['id']})
     response = get(url, u)
     assert 'related' in response.data
@@ -69,8 +68,8 @@ def test_encrypted_subfields(get, post, user, organization):
             notification_configuration=dict(account_sid="dummy", account_token="shouldhide", from_number="+19999999999", to_numbers=["9998887777"]),
         ),
         u,
+        expect=201,
     )
-    assert response.status_code == 201
     notification_template_actual = NotificationTemplate.objects.get(id=response.data['id'])
     url = reverse('api:notification_template_detail', kwargs={'pk': response.data['id']})
     response = get(url, u)
@@ -96,8 +95,8 @@ def test_inherited_notification_templates(get, post, user, organization, project
                 notification_configuration=dict(url="http://localhost", disable_ssl_verification=False, headers={"Test": "Header"}),
             ),
             u,
+            expect=201,
         )
-        assert response.status_code == 201
         notification_templates.append(response.data['id'])
     i = Inventory.objects.create(name='test', organization=organization)
     i.save()
@@ -122,8 +121,7 @@ def test_disallow_delete_when_notifications_pending(delete, user, notification_t
     u = user('superuser', True)
     url = reverse('api:notification_template_detail', kwargs={'pk': notification_template.id})
     Notification.objects.create(notification_template=notification_template, status='pending')
-    response = delete(url, user=u)
-    assert response.status_code == 405
+    delete(url, user=u, expect=405)
 
 
 @pytest.mark.django_db
@@ -133,9 +131,8 @@ def test_notification_template_list_includes_notification_errors(get, user, noti
     Notification.objects.create(notification_template=notification_template, status='successful')
     url = reverse('api:notification_template_list')
     u = user('superuser', True)
-    response = get(url, user=u)
+    response = get(url, user=u, expect=200)
 
-    assert response.status_code == 200
     notifications = response.data['results'][0]['summary_fields']['recent_notifications']
     assert len(notifications) == 3
     statuses = [n['status'] for n in notifications]
@@ -163,8 +160,8 @@ def test_custom_environment_injection(post, user, organization):
             notification_configuration=dict(url="https://example.org", disable_ssl_verification=False, http_method="POST", headers={"Test": "Header"}),
         ),
         u,
+        expect=201,
     )
-    assert response.status_code == 201
     template = NotificationTemplate.objects.get(pk=response.data['id'])
     with pytest.raises(ConnectionError), override_settings(AWX_TASK_ENV={'HTTPS_PROXY': '192.168.50.100:1234'}), mock.patch.object(
         HTTPAdapter, 'send'

awx/main/tests/functional/test_projects.py

@@ -4,7 +4,7 @@ from unittest import mock  # noqa
 import pytest
 
 from awx.api.versioning import reverse
-from awx.main.models import Project
+from awx.main.models import Project, JobTemplate
 
 from django.core.exceptions import ValidationError
 
@@ -451,3 +451,19 @@ def test_project_list_ordering_with_duplicate_names(get, order_by, organization_
         results = get(reverse('api:project_list'), objects.superusers.admin, QUERY_STRING='order_by=%s' % order_by).data['results']
         project_ids[x] = [proj['id'] for proj in results]
     assert project_ids[0] == project_ids[1] == project_ids[2] == [1, 2, 3, 4, 5]
+
+
+@pytest.mark.django_db
+def test_project_failed_update(post, project, admin, inventory):
+    """Test to ensure failed projects with update on launch will create launch rather than error"""
+    jt = JobTemplate.objects.create(project=project, inventory=inventory)
+    # set project to update on launch and set status to failed
+    project.update_fields(scm_update_on_launch=True)
+    project.update()
+    project.project_updates.last().update_fields(status='failed')
+    response = post(reverse('api:job_template_launch', kwargs={'pk': jt.pk}), user=admin, expect=201)
+    assert response.status_code == 201
+    # set project to not update on launch and validate still 400's
+    project.update_fields(scm_update_on_launch=False)
+    response = post(reverse('api:job_template_launch', kwargs={'pk': jt.pk}), user=admin, expect=400)
+    assert response.status_code == 400

awx/main/tests/unit/api/test_logger.py

@@ -47,7 +47,7 @@ data_loggly = {
             '\n'.join(
                 [
                     'template(name="awx" type="string" string="%rawmsg-after-pri%")\nmodule(load="omhttp")',
-                    'action(type="omhttp" server="logs-01.loggly.com" serverport="80" usehttps="off" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" errorfile="/var/log/tower/rsyslog.err" restpath="inputs/1fd38090-2af1-4e1e-8d80-492899da0f71/tag/http/")',  # noqa
+                    'action(type="omhttp" server="logs-01.loggly.com" serverport="80" usehttps="off" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" queue.spoolDirectory="/var/lib/awx" queue.filename="awx-external-logger-action-queue" queue.maxdiskspace="1g" queue.type="LinkedList" queue.saveOnShutdown="on" errorfile="/var/log/tower/rsyslog.err" restpath="inputs/1fd38090-2af1-4e1e-8d80-492899da0f71/tag/http/")',  # noqa
                 ]
             ),
         ),
@@ -89,7 +89,7 @@ data_loggly = {
             '\n'.join(
                 [
                     'template(name="awx" type="string" string="%rawmsg-after-pri%")\nmodule(load="omhttp")',
-                    'action(type="omhttp" server="yoursplunk" serverport="443" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
+                    'action(type="omhttp" server="yoursplunk" serverport="443" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" queue.spoolDirectory="/var/lib/awx" queue.filename="awx-external-logger-action-queue" queue.maxdiskspace="1g" queue.type="LinkedList" queue.saveOnShutdown="on" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
                 ]
             ),
         ),
@@ -103,7 +103,7 @@ data_loggly = {
             '\n'.join(
                 [
                     'template(name="awx" type="string" string="%rawmsg-after-pri%")\nmodule(load="omhttp")',
-                    'action(type="omhttp" server="yoursplunk" serverport="80" usehttps="off" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
+                    'action(type="omhttp" server="yoursplunk" serverport="80" usehttps="off" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" queue.spoolDirectory="/var/lib/awx" queue.filename="awx-external-logger-action-queue" queue.maxdiskspace="1g" queue.type="LinkedList" queue.saveOnShutdown="on" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
                 ]
             ),
         ),
@@ -117,7 +117,7 @@ data_loggly = {
             '\n'.join(
                 [
                     'template(name="awx" type="string" string="%rawmsg-after-pri%")\nmodule(load="omhttp")',
-                    'action(type="omhttp" server="yoursplunk" serverport="8088" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
+                    'action(type="omhttp" server="yoursplunk" serverport="8088" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" queue.spoolDirectory="/var/lib/awx" queue.filename="awx-external-logger-action-queue" queue.maxdiskspace="1g" queue.type="LinkedList" queue.saveOnShutdown="on" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
                 ]
             ),
         ),
@@ -131,7 +131,7 @@ data_loggly = {
             '\n'.join(
                 [
                     'template(name="awx" type="string" string="%rawmsg-after-pri%")\nmodule(load="omhttp")',
-                    'action(type="omhttp" server="yoursplunk" serverport="8088" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
+                    'action(type="omhttp" server="yoursplunk" serverport="8088" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" queue.spoolDirectory="/var/lib/awx" queue.filename="awx-external-logger-action-queue" queue.maxdiskspace="1g" queue.type="LinkedList" queue.saveOnShutdown="on" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
                 ]
             ),
         ),
@@ -145,7 +145,7 @@ data_loggly = {
             '\n'.join(
                 [
                     'template(name="awx" type="string" string="%rawmsg-after-pri%")\nmodule(load="omhttp")',
-                    'action(type="omhttp" server="yoursplunk.org" serverport="8088" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
+                    'action(type="omhttp" server="yoursplunk.org" serverport="8088" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" queue.spoolDirectory="/var/lib/awx" queue.filename="awx-external-logger-action-queue" queue.maxdiskspace="1g" queue.type="LinkedList" queue.saveOnShutdown="on" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
                 ]
             ),
         ),
@@ -159,7 +159,7 @@ data_loggly = {
             '\n'.join(
                 [
                     'template(name="awx" type="string" string="%rawmsg-after-pri%")\nmodule(load="omhttp")',
-                    'action(type="omhttp" server="yoursplunk.org" serverport="8088" usehttps="off" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
+                    'action(type="omhttp" server="yoursplunk.org" serverport="8088" usehttps="off" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" queue.spoolDirectory="/var/lib/awx" queue.filename="awx-external-logger-action-queue" queue.maxdiskspace="1g" queue.type="LinkedList" queue.saveOnShutdown="on" errorfile="/var/log/tower/rsyslog.err" restpath="services/collector/event")',  # noqa
                 ]
             ),
         ),
@@ -173,7 +173,7 @@ data_loggly = {
             '\n'.join(
                 [
                     'template(name="awx" type="string" string="%rawmsg-after-pri%")\nmodule(load="omhttp")',
-                    'action(type="omhttp" server="endpoint5.collection.us2.sumologic.com" serverport="443" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" errorfile="/var/log/tower/rsyslog.err" restpath="receiver/v1/http/ZaVnC4dhaV0qoiETY0MrM3wwLoDgO1jFgjOxE6-39qokkj3LGtOroZ8wNaN2M6DtgYrJZsmSi4-36_Up5TbbN_8hosYonLKHSSOSKY845LuLZBCBwStrHQ==")',  # noqa
+                    'action(type="omhttp" server="endpoint5.collection.us2.sumologic.com" serverport="443" usehttps="on" allowunsignedcerts="off" skipverifyhost="off" action.resumeRetryCount="-1" template="awx" action.resumeInterval="5" queue.spoolDirectory="/var/lib/awx" queue.filename="awx-external-logger-action-queue" queue.maxdiskspace="1g" queue.type="LinkedList" queue.saveOnShutdown="on" errorfile="/var/log/tower/rsyslog.err" restpath="receiver/v1/http/ZaVnC4dhaV0qoiETY0MrM3wwLoDgO1jFgjOxE6-39qokkj3LGtOroZ8wNaN2M6DtgYrJZsmSi4-36_Up5TbbN_8hosYonLKHSSOSKY845LuLZBCBwStrHQ==")',  # noqa
                 ]
             ),
         ),

awx/main/utils/common.py

@@ -90,6 +90,7 @@ __all__ = [
     'get_event_partition_epoch',
     'cleanup_new_process',
     'log_excess_runtime',
+    'unified_job_class_to_event_table_name',
 ]
 
 
@@ -1219,3 +1220,7 @@ def log_excess_runtime(func_logger, cutoff=5.0, debug_cutoff=5.0, msg=None, add_
         return _new_func
 
     return log_excess_runtime_decorator
+
+
+def unified_job_class_to_event_table_name(job_class):
+    return f'main_{job_class().event_class.__name__.lower()}'

awx/main/utils/external_logging.py

@@ -17,7 +17,8 @@ def construct_rsyslog_conf_template(settings=settings):
     port = getattr(settings, 'LOG_AGGREGATOR_PORT', '')
     protocol = getattr(settings, 'LOG_AGGREGATOR_PROTOCOL', '')
     timeout = getattr(settings, 'LOG_AGGREGATOR_TCP_TIMEOUT', 5)
-    max_disk_space = getattr(settings, 'LOG_AGGREGATOR_MAX_DISK_USAGE_GB', 1)
+    max_disk_space_main_queue = getattr(settings, 'LOG_AGGREGATOR_MAX_DISK_USAGE_GB', 1)
+    max_disk_space_action_queue = getattr(settings, 'LOG_AGGREGATOR_ACTION_MAX_DISK_USAGE_GB', 1)
     spool_directory = getattr(settings, 'LOG_AGGREGATOR_MAX_DISK_USAGE_PATH', '/var/lib/awx').rstrip('/')
     error_log_file = getattr(settings, 'LOG_AGGREGATOR_RSYSLOGD_ERROR_LOG_FILE', '')
 
@@ -32,7 +33,7 @@ def construct_rsyslog_conf_template(settings=settings):
             '$WorkDirectory /var/lib/awx/rsyslog',
             f'$MaxMessageSize {max_bytes}',
             '$IncludeConfig /var/lib/awx/rsyslog/conf.d/*.conf',
-            f'main_queue(queue.spoolDirectory="{spool_directory}" queue.maxdiskspace="{max_disk_space}g" queue.type="Disk" queue.filename="awx-external-logger-backlog")',  # noqa
+            f'main_queue(queue.spoolDirectory="{spool_directory}" queue.maxdiskspace="{max_disk_space_main_queue}g" queue.type="Disk" queue.filename="awx-external-logger-backlog")',  # noqa
             'module(load="imuxsock" SysSock.Use="off")',
             'input(type="imuxsock" Socket="' + settings.LOGGING['handlers']['external_logger']['address'] + '" unlink="on" RateLimit.Burst="0")',
             'template(name="awx" type="string" string="%rawmsg-after-pri%")',
@@ -78,6 +79,11 @@ def construct_rsyslog_conf_template(settings=settings):
             'action.resumeRetryCount="-1"',
             'template="awx"',
             f'action.resumeInterval="{timeout}"',
+            f'queue.spoolDirectory="{spool_directory}"',
+            'queue.filename="awx-external-logger-action-queue"',
+            f'queue.maxdiskspace="{max_disk_space_action_queue}g"',
+            'queue.type="LinkedList"',
+            'queue.saveOnShutdown="on"',
         ]
         if error_log_file:
             params.append(f'errorfile="{error_log_file}"')

awx/main/utils/handlers.py

@@ -97,8 +97,6 @@ class SpecialInventoryHandler(logging.Handler):
         self.event_handler(dispatch_data)
 
 
-ColorHandler = logging.StreamHandler
-
 if settings.COLOR_LOGS is True:
     try:
         from logutils.colorize import ColorizingStreamHandler
@@ -133,3 +131,5 @@ if settings.COLOR_LOGS is True:
     except ImportError:
         # logutils is only used for colored logs in the dev environment
         pass
+else:
+    ColorHandler = logging.StreamHandler

awx/main/utils/licensing.py

@@ -175,7 +175,12 @@ class Licenser(object):
             license.setdefault('pool_id', sub['pool']['id'])
             license.setdefault('product_name', sub['pool']['productName'])
             license.setdefault('valid_key', True)
-            license.setdefault('license_type', 'enterprise')
+            if sub['pool']['productId'].startswith('S'):
+                license.setdefault('trial', True)
+                license.setdefault('license_type', 'trial')
+            else:
+                license.setdefault('trial', False)
+                license.setdefault('license_type', 'enterprise')
             license.setdefault('satellite', False)
             # Use the nearest end date
             endDate = parse_date(sub['endDate'])
@@ -287,7 +292,7 @@ class Licenser(object):
                     license['productId'] = sub['product_id']
                     license['quantity'] = int(sub['quantity'])
                     license['support_level'] = sub['support_level']
-                    license['usage'] = sub['usage']
+                    license['usage'] = sub.get('usage')
                     license['subscription_name'] = sub['name']
                     license['subscriptionId'] = sub['subscription_id']
                     license['accountNumber'] = sub['account_number']

awx/main/wsrelay.py

@@ -12,7 +12,7 @@ from channels.layers import get_channel_layer
 from django.conf import settings
 from django.apps import apps
 
-import asyncpg
+import psycopg
 
 from awx.main.analytics.broadcast_websocket import (
     RelayWebsocketStats,
@@ -209,53 +209,49 @@ class WebSocketRelayManager(object):
         # hostname -> ip
         self.known_hosts: Dict[str, str] = dict()
 
-    async def on_heartbeet(self, conn, pid, channel, payload):
-        try:
-            if not payload or channel != "web_heartbeet":
-                return
-
+    async def on_ws_heartbeat(self, conn):
+        await conn.execute("LISTEN web_ws_heartbeat")
+        async for notif in conn.notifies():
+            if notif is None:
+                continue
             try:
-                payload = json.loads(payload)
-            except json.JSONDecodeError:
-                logmsg = "Failed to decode message from pg_notify channel `web_heartbeet`"
-                if logger.isEnabledFor(logging.DEBUG):
-                    logmsg = "{} {}".format(logmsg, payload)
-                    logger.warning(logmsg)
-                return
-
-            # Skip if the message comes from the same host we are running on
-            # In this case, we'll be sharing a redis, no need to relay.
-            if payload.get("hostname") == self.local_hostname:
-                return
-
-            if payload.get("action") == "online":
-                hostname = payload.get("hostname")
-                ip = payload.get("ip")
-                if ip is None:
-                    # If we don't get an IP, just try the hostname, maybe it resolves
-                    ip = hostname
-                if ip is None:
-                    logger.warning(f"Received invalid online heartbeet, missing hostname and ip: {payload}")
+                if not notif.payload or notif.channel != "web_ws_heartbeat":
                     return
-                self.known_hosts[hostname] = ip
-                logger.debug(f"Web host {hostname} ({ip}) online heartbeat received.")
-            elif payload.get("action") == "offline":
-                hostname = payload.get("hostname")
-                ip = payload.get("ip")
-                if ip is None:
-                    # If we don't get an IP, just try the hostname, maybe it resolves
-                    ip = hostname
-                if ip is None:
-                    logger.warning(f"Received invalid offline heartbeet, missing hostname and ip: {payload}")
-                    return
-                self.cleanup_offline_host(ip)
-                logger.debug(f"Web host {hostname} ({ip}) offline heartbeat received.")
-        except Exception as e:
-            # This catch-all is the same as the one above. asyncio will eat the exception
-            # but we want to know about it.
-            logger.exception(f"on_heartbeet exception: {e}")
 
-    def cleanup_offline_host(self, hostname):
+                try:
+                    payload = json.loads(notif.payload)
+                except json.JSONDecodeError:
+                    logmsg = "Failed to decode message from pg_notify channel `web_ws_heartbeat`"
+                    if logger.isEnabledFor(logging.DEBUG):
+                        logmsg = "{} {}".format(logmsg, payload)
+                        logger.warning(logmsg)
+                    return
+
+                # Skip if the message comes from the same host we are running on
+                # In this case, we'll be sharing a redis, no need to relay.
+                if payload.get("hostname") == self.local_hostname:
+                    return
+
+                action = payload.get("action")
+
+                if action in ("online", "offline"):
+                    hostname = payload.get("hostname")
+                    ip = payload.get("ip") or hostname  # try back to hostname if ip isn't supplied
+                    if ip is None:
+                        logger.warning(f"Received invalid {action} ws_heartbeat, missing hostname and ip: {payload}")
+                        return
+                    logger.debug(f"Web host {hostname} ({ip}) {action} heartbeat received.")
+
+                if action == "online":
+                    self.known_hosts[hostname] = ip
+                elif action == "offline":
+                    await self.cleanup_offline_host(hostname)
+            except Exception as e:
+                # This catch-all is the same as the one above. asyncio will eat the exception
+                # but we want to know about it.
+                logger.exception(f"on_ws_heartbeat exception: {e}")
+
+    async def cleanup_offline_host(self, hostname):
         """
         Given a hostname, try to cancel its task/connection and remove it from
         the list of hosts we know about.
@@ -264,6 +260,19 @@ class WebSocketRelayManager(object):
         """
         if hostname in self.relay_connections:
             self.relay_connections[hostname].cancel()
+
+            # Wait for the task to actually run its cancel/completion logic
+            # otherwise it might get GC'd too early when we del it below.
+            # Being GC'd too early could generate a scary message in logs:
+            # "Task was destroyed but it is pending!"
+            try:
+                await asyncio.wait_for(self.relay_connections[hostname].async_task, timeout=10)
+            except asyncio.TimeoutError:
+                logger.warning(f"Tried to cancel relay connection for {hostname} but it timed out during cleanup.")
+            except asyncio.CancelledError:
+                # Handle the case where the task was already cancelled by the time we got here.
+                pass
+
             del self.relay_connections[hostname]
 
         if hostname in self.known_hosts:
@@ -282,16 +291,16 @@ class WebSocketRelayManager(object):
 
         # Set up a pg_notify consumer for allowing web nodes to "provision" and "deprovision" themselves gracefully.
         database_conf = settings.DATABASES['default']
-        async_conn = await asyncpg.connect(
-            database=database_conf['NAME'],
+        async_conn = await psycopg.AsyncConnection.connect(
+            dbname=database_conf['NAME'],
             host=database_conf['HOST'],
             user=database_conf['USER'],
             password=database_conf['PASSWORD'],
             port=database_conf['PORT'],
-            # We cannot include these because asyncpg doesn't allow all the options that psycopg does.
-            # **database_conf.get("OPTIONS", {}),
+            **database_conf.get("OPTIONS", {}),
         )
-        await async_conn.add_listener("web_heartbeet", self.on_heartbeet)
+        await async_conn.set_autocommit(True)
+        event_loop.create_task(self.on_ws_heartbeat(async_conn))
 
         # Establishes a websocket connection to /websocket/relay on all API servers
         while True:
@@ -318,13 +327,11 @@ class WebSocketRelayManager(object):
 
             if deleted_remote_hosts:
                 logger.info(f"Removing {deleted_remote_hosts} from websocket broadcast list")
+                await asyncio.gather(self.cleanup_offline_host(h) for h in deleted_remote_hosts)
 
             if new_remote_hosts:
                 logger.info(f"Adding {new_remote_hosts} to websocket broadcast list")
 
-            for h in deleted_remote_hosts:
-                self.cleanup_offline_host(h)
-
             for h in new_remote_hosts:
                 stats = self.stats_mgr.new_remote_host_stats(h)
                 relay_connection = WebsocketRelayConnection(name=self.local_hostname, stats=stats, remote_host=self.known_hosts[h])

awx/playbooks/project_update.yml

@@ -189,11 +189,12 @@
   connection: local
   name: Install content with ansible-galaxy command if necessary
   vars:
-    galaxy_task_env:  # configure in settings
-    additional_collections_env:
-      # These environment variables are used for installing collections, in addition to galaxy_task_env
-      # setting the collections paths silences warnings
+    galaxy_task_env:  # configured in settings
+    # additional_galaxy_env contains environment variables are used for installing roles and collections and will take precedence over items in galaxy_task_env
+    additional_galaxy_env:
+      # These paths control where ansible-galaxy installs collections and roles on top the filesystem
       ANSIBLE_COLLECTIONS_PATHS: "{{ projects_root }}/.__awx_cache/{{ local_path }}/stage/requirements_collections"
+      ANSIBLE_ROLES_PATH: "{{ projects_root }}/.__awx_cache/{{ local_path }}/stage/requirements_roles"
       # Put the local tmp directory in same volume as collection destination
       # otherwise, files cannot be moved accross volumes and will cause error
       ANSIBLE_LOCAL_TEMP: "{{ projects_root }}/.__awx_cache/{{ local_path }}/stage/tmp"
@@ -212,40 +213,53 @@
         - name: End play due to disabled content sync
           ansible.builtin.meta: end_play
 
-    - name: Fetch galaxy roles from requirements.(yml/yaml)
-      ansible.builtin.command: >
-        ansible-galaxy role install -r {{ item }}
-        --roles-path {{ projects_root }}/.__awx_cache/{{ local_path }}/stage/requirements_roles
-        {{ ' -' + 'v' * ansible_verbosity if ansible_verbosity else '' }}
-      args:
-        chdir: "{{ project_path | quote }}"
-      register: galaxy_result
-      with_fileglob:
-        - "{{ project_path | quote }}/roles/requirements.yaml"
-        - "{{ project_path | quote }}/roles/requirements.yml"
-      changed_when: "'was installed successfully' in galaxy_result.stdout"
-      environment: "{{ galaxy_task_env }}"
-      when: roles_enabled | bool
-      tags:
-        - install_roles
+    - block:
+        - name: Fetch galaxy roles from roles/requirements.(yml/yaml)
+          ansible.builtin.command:
+            cmd: "ansible-galaxy role install -r {{ item }} {{ verbosity }}"
+          register: galaxy_result
+          with_fileglob:
+            - "{{ project_path | quote }}/roles/requirements.yaml"
+            - "{{ project_path | quote }}/roles/requirements.yml"
+          changed_when: "'was installed successfully' in galaxy_result.stdout"
+          when: roles_enabled | bool
+          tags:
+            - install_roles
 
-    - name: Fetch galaxy collections from collections/requirements.(yml/yaml)
-      ansible.builtin.command: >
-        ansible-galaxy collection install -r {{ item }}
-        --collections-path {{ projects_root }}/.__awx_cache/{{ local_path }}/stage/requirements_collections
-        {{ ' -' + 'v' * ansible_verbosity if ansible_verbosity else '' }}
-      args:
-        chdir: "{{ project_path | quote }}"
-      register: galaxy_collection_result
-      with_fileglob:
-        - "{{ project_path | quote }}/collections/requirements.yaml"
-        - "{{ project_path | quote }}/collections/requirements.yml"
-        - "{{ project_path | quote }}/requirements.yaml"
-        - "{{ project_path | quote }}/requirements.yml"
-      changed_when: "'Installing ' in galaxy_collection_result.stdout"
-      environment: "{{ additional_collections_env | combine(galaxy_task_env) }}"
-      when:
-        - "ansible_version.full is version_compare('2.9', '>=')"
-        - collections_enabled | bool
-      tags:
-        - install_collections
+        - name: Fetch galaxy collections from collections/requirements.(yml/yaml)
+          ansible.builtin.command:
+            cmd: "ansible-galaxy collection install -r {{ item }} {{ verbosity }}"
+          register: galaxy_collection_result
+          with_fileglob:
+            - "{{ project_path | quote }}/collections/requirements.yaml"
+            - "{{ project_path | quote }}/collections/requirements.yml"
+          changed_when: "'Nothing to do.' not in galaxy_collection_result.stdout"
+          when:
+            - "ansible_version.full is version_compare('2.9', '>=')"
+            - collections_enabled | bool
+          tags:
+            - install_collections
+
+        - name: Fetch galaxy roles and collections from requirements.(yml/yaml)
+          ansible.builtin.command:
+            cmd: "ansible-galaxy install -r {{ item }} {{ verbosity }}"
+          register: galaxy_combined_result
+          with_fileglob:
+            - "{{ project_path | quote }}/requirements.yaml"
+            - "{{ project_path | quote }}/requirements.yml"
+          changed_when: "'Nothing to do.' not in galaxy_combined_result.stdout"
+          when:
+            - "ansible_version.full is version_compare('2.10', '>=')"
+            - collections_enabled | bool
+            - roles_enabled | bool
+          tags:
+            - install_collections
+            - install_roles
+      module_defaults:
+        ansible.builtin.command:
+          chdir: "{{ project_path | quote }}"
+
+      # We combine our additional_galaxy_env into galaxy_task_env so that our values are preferred over anything a user would set
+      environment: "{{ galaxy_task_env | combine(additional_galaxy_env) }}"
+      vars:
+        verbosity: "{{ (ansible_verbosity) | ternary('-'+'v'*ansible_verbosity, '') }}"