Document, implement, and test remaining indirect host audit fields

awx/main/migrations/0201_indirect_managed_node_audit.py

@@ -0,0 +1,85 @@
+# Generated by Django 4.2.16 on 2025-01-29 20:13
+
+from django.db import migrations, models
+import django.db.models.deletion
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('main', '0200_delete_token_cleanup_job'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='job',
+            name='event_queries_processed',
+            field=models.BooleanField(default=False, help_text='Events of this job have been queried for indirect host information'),
+        ),
+        migrations.CreateModel(
+            name='EventQuery',
+            fields=[
+                ('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('fqcn', models.CharField(help_text='Fully-qualified collection name.', max_length=255)),
+                ('collection_version', models.CharField(help_text='Version of the collection this data applies to.', max_length=32)),
+                ('event_query', models.JSONField(default=dict, help_text='The extensions/audit/event_query.yml file content scraped from the collection.')),
+            ],
+            options={
+                'unique_together': {('fqcn', 'collection_version')},
+            },
+        ),
+        migrations.CreateModel(
+            name='IndirectManagedNodeAudit',
+            fields=[
+                ('id', models.AutoField(auto_created=True, primary_key=True, serialize=False, verbose_name='ID')),
+                ('created', models.DateTimeField(auto_now_add=True)),
+                ('name', models.CharField(help_text='The Ansible name of the host that this audit record is for.', max_length=255)),
+                ('canonical_facts', models.JSONField(default=dict, help_text='Facts about the host that will be used for managed node deduplication.')),
+                ('facts', models.JSONField(default=dict, help_text='Non canonical facts having additional info about the managed node.')),
+                ('events', models.JSONField(default=list, help_text='List of fully-qualified names of modules that ran against the host in the job.')),
+                ('count', models.PositiveIntegerField(default=0, help_text='Counter of how many times registered modules were invoked on the host.')),
+                (
+                    'host',
+                    models.ForeignKey(
+                        help_text='The host this audit record is for.',
+                        null=True,
+                        on_delete=django.db.models.deletion.DO_NOTHING,
+                        related_name='host_indirect_host_audits',
+                        to='main.host',
+                    ),
+                ),
+                (
+                    'inventory',
+                    models.ForeignKey(
+                        help_text='The inventory the related job ran against, and which the related host is in.',
+                        null=True,
+                        on_delete=django.db.models.deletion.DO_NOTHING,
+                        related_name='inventory_indirect_host_audits',
+                        to='main.inventory',
+                    ),
+                ),
+                (
+                    'job',
+                    models.ForeignKey(
+                        editable=False,
+                        help_text='Data saved in this record only applies to this specified job.',
+                        on_delete=django.db.models.deletion.DO_NOTHING,
+                        related_name='job_indirect_host_audits',
+                        to='main.job',
+                    ),
+                ),
+                (
+                    'organization',
+                    models.ForeignKey(
+                        help_text='Applicable organization, inferred from the related job.',
+                        on_delete=django.db.models.deletion.DO_NOTHING,
+                        related_name='organization_indirect_host_audits',
+                        to='main.organization',
+                    ),
+                ),
+            ],
+            options={
+                'unique_together': {('name', 'job')},
+            },
+        ),
+    ]

awx/main/models/event_query.py

@@ -1,5 +1,6 @@
 from django.core.exceptions import ValidationError
 from django.db import models
+from django.utils.translation import gettext_lazy as _
 
 from awx.main.models import BaseModel
 
@@ -14,9 +15,9 @@ class Meta:
         app_label = 'main'
         unique_together = ['fqcn', 'collection_version']
 
-    fqcn = models.CharField(max_length=255)
-    collection_version = models.CharField(max_length=32)
-    event_query = models.JSONField(default=dict)
+    fqcn = models.CharField(max_length=255, help_text=_('Fully-qualified collection name.'))
+    collection_version = models.CharField(max_length=32, help_text=_('Version of the collection this data applies to.'))
+    event_query = models.JSONField(default=dict, help_text=_('The extensions/audit/event_query.yml file content scraped from the collection.'))
 
     def validate_unique(self, exclude=None):
         try:

awx/main/models/indirect_managed_node_audit.py

@@ -2,6 +2,8 @@
 from django.db.models.fields import DateTimeField, CharField, PositiveIntegerField
 from django.db.models.fields.json import JSONField
 from django.db.models.fields.related import ForeignKey
+from django.utils.translation import gettext_lazy as _
+
 from awx.main.models import BaseModel
 
 
@@ -21,34 +23,32 @@ class Meta:
         related_name='job_indirect_host_audits',
         on_delete=DO_NOTHING,
         editable=False,
+        help_text=_('Data saved in this record only applies to this specified job.'),
     )
 
     organization = ForeignKey(
         'Organization',
         related_name='organization_indirect_host_audits',
         on_delete=DO_NOTHING,
+        help_text=_('Applicable organization, inferred from the related job.'),
     )
 
     inventory = ForeignKey(
         'Inventory',
         related_name='inventory_indirect_host_audits',
         null=True,
         on_delete=DO_NOTHING,
+        help_text=_('The inventory the related job ran against, and which the related host is in.'),
     )
 
-    host = ForeignKey(
-        'Host',
-        related_name='host_indirect_host_audits',
-        null=True,
-        on_delete=DO_NOTHING,
-    )
+    host = ForeignKey('Host', related_name='host_indirect_host_audits', null=True, on_delete=DO_NOTHING, help_text=_('The host this audit record is for.'))
 
-    name = CharField(max_length=255)
+    name = CharField(max_length=255, help_text=_('The Ansible name of the host that this audit record is for.'))
 
-    canonical_facts = JSONField(default=dict)
+    canonical_facts = JSONField(default=dict, help_text=_('Facts about the host that will be used for managed node deduplication.'))
 
-    facts = JSONField(default=dict)
+    facts = JSONField(default=dict, help_text=_('Non canonical facts having additional info about the managed node.'))
 
-    events = JSONField(default=list)
+    events = JSONField(default=list, help_text=_('List of fully-qualified names of modules that ran against the host in the job.'))
 
-    count = PositiveIntegerField(default=0)
+    count = PositiveIntegerField(default=0, help_text=_('Counter of how many times registered modules were invoked on the host.'))

awx/main/tasks/host_indirect.py

@@ -1,4 +1,5 @@
 import logging
+from typing import Tuple, Union
 
 import yaml
 
@@ -13,26 +14,72 @@
 logger = logging.getLogger(__name__)
 
 
+class UnhashableFacts(RuntimeError):
+    pass
+
+
+def get_hashable_form(python_dict: Union[dict, list, int, float, str, bool]) -> Tuple[Union[Tuple, dict, int, float]]:
+    "Given a dictionary of JSON types, return something that can be hashed and is the same data"
+    if isinstance(python_dict, (int, float, str, bool)):
+        return python_dict  # return scalars as-is
+    if isinstance(python_dict, dict):
+        # Can't hash? Make it a tuple. Can't hash the tuples in the tuple? We'll make tuples out of them too.
+        return tuple(sorted(((get_hashable_form(k), get_hashable_form(v)) for k, v in python_dict.items())))
+    elif isinstance(python_dict, (list, tuple)):
+        return tuple(python_dict)
+    raise UnhashableFacts(f'Cannonical facts contains a {type(python_dict)} type which can not be hashed.')
+
+
 def build_indirect_host_data(job, job_event_queries: dict[str, str]) -> list[IndirectManagedNodeAudit]:
-    results = []
+    results = {}
     compiled_jq_expressions = {}  # Cache for compiled jq expressions
     facts_missing_logged = False
+    unhashable_facts_logged = False
+    print(f'using event queries {job_event_queries}')
     for event in job.job_events.filter(task__in=job_event_queries.keys()).iterator():
+        print(f'inspecting event {event}')
         if 'res' not in event.event_data:
             continue
         jq_str_for_event = job_event_queries[event.task]
         if jq_str_for_event not in compiled_jq_expressions:
             compiled_jq_expressions[event.task] = jq.compile(jq_str_for_event)
         compiled_jq = compiled_jq_expressions[event.task]
         for data in compiled_jq.input(event.event_data['res']).all():
+
+            # From the JQ result, get index information about this record
             if not data.get('canonical_facts'):
                 if not facts_missing_logged:
                     logger.error(f'jq output missing canonical_facts for module {event.task} on event {event.id} using jq:{jq_str_for_event}')
                 continue
             canonical_facts = data['canonical_facts']
+            try:
+                hashable_facts = get_hashable_form(canonical_facts)
+            except UnhashableFacts:
+                if not unhashable_facts_logged:
+                    logger.info(f'Could not hash canonical_facts {canonical_facts}, skipping')
+                    unhashable_facts_logged = True
+                continue
+
+            # Obtain the record based on the hashable canonical_facts now determined
             facts = data.get('facts')
-            results.append(IndirectManagedNodeAudit(canonical_facts=canonical_facts, facts=facts, job=job, organization=job.organization))
-    return results
+            if hashable_facts in results:
+                audit_record = results[hashable_facts]
+            else:
+                audit_record = IndirectManagedNodeAudit(
+                    canonical_facts=canonical_facts,
+                    facts=facts,
+                    job=job,
+                    organization=job.organization,
+                    name=event.host_name,
+                )
+                results[hashable_facts] = audit_record
+
+            # Increment rolling count fields
+            if event.task not in audit_record.events:
+                audit_record.events.append(event.task)
+            audit_record.count += 1
+
+    return list(results.values())
 
 
 def fetch_job_event_query(job) -> dict[str, str]:
@@ -59,3 +106,5 @@ def save_indirect_host_entries(job_id):
     job_event_queries = fetch_job_event_query(job)
     records = build_indirect_host_data(job, job_event_queries)
     IndirectManagedNodeAudit.objects.bulk_create(records)
+    job.event_queries_processed = True
+    job.save(update_fields=['event_queries_processed'])