awx/awx/main/models/ha.py
Matthew Jones 70bf78e29f
Apply capacity algorithm changes
* This also adds fields to the instance view for tracking cpu and
  memory usage as well as information on what the capacity ranges are
* Also adds a flag for enabling/disabling instances which removes them
  from all queues and has them stop processing new work
* The capacity is now based almost exclusively on some value relative
  to forks
* capacity_adjustment allows you to commit an instance to a certain
  amount of forks, cpu focused or memory focused
* Each job run adds a single fork overhead (that's the reasoning
  behind the +1)
2018-02-01 16:57:09 -05:00

230 lines
7.9 KiB
Python

# Copyright (c) 2015 Ansible, Inc.
# All Rights Reserved.
from decimal import Decimal
from django.db import models, connection
from django.db.models.signals import post_save, post_delete
from django.dispatch import receiver
from django.utils.translation import ugettext_lazy as _
from django.conf import settings
from django.utils.timezone import now, timedelta
from solo.models import SingletonModel
from awx import __version__ as awx_application_version
from awx.api.versioning import reverse
from awx.main.managers import InstanceManager, InstanceGroupManager
from awx.main.fields import JSONField
from awx.main.models.inventory import InventoryUpdate
from awx.main.models.jobs import Job
from awx.main.models.projects import ProjectUpdate
from awx.main.models.unified_jobs import UnifiedJob
from awx.main.utils import get_cpu_capacity, get_mem_capacity, get_system_task_capacity
__all__ = ('Instance', 'InstanceGroup', 'JobOrigin', 'TowerScheduleState',)
class Instance(models.Model):
"""A model representing an AWX instance running against this database."""
objects = InstanceManager()
uuid = models.CharField(max_length=40)
hostname = models.CharField(max_length=250, unique=True)
created = models.DateTimeField(auto_now_add=True)
modified = models.DateTimeField(auto_now=True)
last_isolated_check = models.DateTimeField(
null=True,
editable=False,
auto_now_add=True
)
version = models.CharField(max_length=24, blank=True)
capacity = models.PositiveIntegerField(
default=100,
editable=False,
)
capacity_adjustment = models.DecimalField(
default=Decimal(1.0),
max_digits=3,
decimal_places=2,
)
enabled = models.BooleanField(
default=True
)
cpu = models.IntegerField(
default=0,
editable=False,
)
memory = models.BigIntegerField(
default=0,
editable=False,
)
cpu_capacity = models.IntegerField(
default=0,
editable=False,
)
mem_capacity = models.IntegerField(
default=0,
editable=False,
)
class Meta:
app_label = 'main'
def get_absolute_url(self, request=None):
return reverse('api:instance_detail', kwargs={'pk': self.pk}, request=request)
@property
def consumed_capacity(self):
return sum(x.task_impact for x in UnifiedJob.objects.filter(execution_node=self.hostname,
status__in=('running', 'waiting')))
@property
def role(self):
# NOTE: TODO: Likely to repurpose this once standalone ramparts are a thing
return "awx"
def is_lost(self, ref_time=None, isolated=False):
if ref_time is None:
ref_time = now()
grace_period = 120
if isolated:
grace_period = settings.AWX_ISOLATED_PERIODIC_CHECK * 2
return self.modified < ref_time - timedelta(seconds=grace_period)
def is_controller(self):
return Instance.objects.filter(rampart_groups__controller__instances=self).exists()
def refresh_capacity(self):
cpu = get_cpu_capacity()
mem = get_mem_capacity()
self.capacity = get_system_task_capacity(self.capacity_adjustment)
self.cpu = cpu[0]
self.memory = mem[0]
self.cpu_capacity = cpu[1]
self.mem_capacity = mem[1]
self.version = awx_application_version
self.save(update_fields=['capacity', 'version', 'modified', 'cpu',
'memory', 'cpu_capacity', 'mem_capacity'])
class InstanceGroup(models.Model):
"""A model representing a Queue/Group of AWX Instances."""
objects = InstanceGroupManager()
name = models.CharField(max_length=250, unique=True)
created = models.DateTimeField(auto_now_add=True)
modified = models.DateTimeField(auto_now=True)
instances = models.ManyToManyField(
'Instance',
related_name='rampart_groups',
editable=False,
help_text=_('Instances that are members of this InstanceGroup'),
)
controller = models.ForeignKey(
'InstanceGroup',
related_name='controlled_groups',
help_text=_('Instance Group to remotely control this group.'),
editable=False,
default=None,
null=True
)
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 = JSONField(
default=[],
blank=True,
help_text=_("List of exact-match Instances that will always be automatically assigned to this group")
)
def get_absolute_url(self, request=None):
return reverse('api:instance_group_detail', kwargs={'pk': self.pk}, request=request)
@property
def capacity(self):
return sum([inst.capacity for inst in self.instances.all()])
class Meta:
app_label = 'main'
class TowerScheduleState(SingletonModel):
schedule_last_run = models.DateTimeField(auto_now_add=True)
class JobOrigin(models.Model):
"""A model representing the relationship between a unified job and
the instance that was responsible for starting that job.
It may be possible that a job has no origin (the common reason for this
being that the job was started on Tower < 2.1 before origins were a thing).
This is fine, and code should be able to handle it. A job with no origin
is always assumed to *not* have the current instance as its origin.
"""
unified_job = models.OneToOneField(UnifiedJob, related_name='job_origin')
instance = models.ForeignKey(Instance)
created = models.DateTimeField(auto_now_add=True)
modified = models.DateTimeField(auto_now=True)
class Meta:
app_label = 'main'
@receiver(post_save, sender=InstanceGroup)
def on_instance_group_saved(sender, instance, created=False, raw=False, **kwargs):
if created:
from awx.main.tasks import apply_cluster_membership_policies
connection.on_commit(lambda: apply_cluster_membership_policies.apply_async())
@receiver(post_save, sender=Instance)
def on_instance_saved(sender, instance, created=False, raw=False, **kwargs):
if created:
from awx.main.tasks import apply_cluster_membership_policies
connection.on_commit(lambda: apply_cluster_membership_policies.apply_async())
@receiver(post_delete, sender=InstanceGroup)
def on_instance_group_deleted(sender, instance, using, **kwargs):
from awx.main.tasks import apply_cluster_membership_policies
connection.on_commit(lambda: apply_cluster_membership_policies.apply_async())
@receiver(post_delete, sender=Instance)
def on_instance_deleted(sender, instance, using, **kwargs):
from awx.main.tasks import apply_cluster_membership_policies
connection.on_commit(lambda: apply_cluster_membership_policies.apply_async())
# Unfortunately, the signal can't just be connected against UnifiedJob; it
# turns out that creating a model's subclass doesn't fire the signal for the
# superclass model.
@receiver(post_save, sender=InventoryUpdate)
@receiver(post_save, sender=Job)
@receiver(post_save, sender=ProjectUpdate)
def on_job_create(sender, instance, created=False, raw=False, **kwargs):
"""When a new job is created, save a record of its origin (the machine
that started the job).
"""
# Sanity check: We only want to create a JobOrigin record in cases where
# we are making a new record, and in normal situations.
#
# In other situations, we simply do nothing.
if raw or not created:
return
# Create the JobOrigin record, which attaches to the current instance
# (which started the job).
job_origin, new = JobOrigin.objects.get_or_create(
instance=Instance.objects.me(),
unified_job=instance,
)