Performance testing and benchmarking (#298)

* Delete all

* Create basic data 1

* Helpers

* Only 10 hosts

* Create jobs

* Create job host summary

* Create events

* Create events partitions

* Batch insert

* Logs errors

* Ruff

* Arguments

* Run_perf script

* Run_perf script - save rollups and json

* Ability to save plain csv rollup data

* Mkdir

* Templates

* Add HostId into events

* Fix job template name in jobs data

* More modules

* Run the whole perf test from one file

* Git ignore out files and add modules

* 3 commands

* Random item names

* Print elapsed time

* Print memory

* Ruff line too long

* Rename folder

* gitignore

* collector performance testing

* run uvx ruff check

* rename run_collector_test.py

* add job_finished

---------

Co-authored-by: MilanPospisil <arkanus@seznam.cz>

Author: ShaiahWren

tools/anonymized_db_perf_data/fill_perf_db_data.py

@@ -112,34 +112,34 @@ def fill_perf_db_data(host_count=10, job_count=5, task_count=50, template_count=
 
 
 def fill_job_data(init_data, job_index):
-    """Create a job using the init_data IDs. Returns (job_id, job_created)."""
+    """Create a job using the init_data IDs. Returns (job_id, job_created, job_finished)."""
     # Randomly select a job template
     templates = init_data['templates']
     template_id = random.choice(list(templates.keys()))
     template_name = templates[template_id]
 
-    job_id, job_created = create_job(
+    job_id, job_created, job_finished = create_job(
         name=template_name,  # Use template name as job name (AWX behavior)
         inventory_id=init_data['inventory_id'],
         project_id=init_data['project_id'],
         org_id=init_data['org_id'],
         job_index=job_index,
         job_template_id=template_id,
     )
-    return job_id, job_created
+    return job_id, job_created, job_finished
 
 
-def fill_jobhostsummary(init_data, job_id):
-    create_job_host_summaries(job_id, init_data['host_count'], unique_suffix=init_data.get('unique_suffix'))
+def fill_jobhostsummary(init_data, job_id, job_created, job_finished):
+    create_job_host_summaries(job_id, init_data['host_count'], job_created, job_finished, unique_suffix=init_data.get('unique_suffix'))
 
 
 def fill_jobevent(init_data, job_id, job_index, job_created):
     create_job_events(job_id, init_data['host_ids'], init_data['task_count'], job_index, job_created, unique_suffix=init_data.get('unique_suffix'))
 
 
 def fill_job(init_data, job_index):
-    job_id, job_created = fill_job_data(init_data, job_index)
-    fill_jobhostsummary(init_data, job_id)
+    job_id, job_created, job_finished = fill_job_data(init_data, job_index)
+    fill_jobhostsummary(init_data, job_id, job_created, job_finished)
     fill_jobevent(init_data, job_id, job_index, job_created)
     return
 

tools/anonymized_db_perf_data/helpers.py

@@ -459,18 +459,20 @@ def create_job(name='Perf Test Job', inventory_id=None, project_id=None, org_id=
     run(sql_job)
     print(f'Created job {job_index}')
 
-    # Return job_id and created timestamp (needed for events)
-    return job_id, created
+    # Return job_id, created timestamp (needed for events), and finished timestamp (needed for job host summaries)
+    return job_id, created, finished
 
 
-def create_job_host_summaries(job_id, host_count, unique_suffix=None):
+def create_job_host_summaries(job_id, host_count, job_created, job_finished, unique_suffix=None):
     """Create job host summaries for all hosts (batch insert).
 
     Host names are generated using the same pattern as create_hosts: host-{i}-{suffix}.example.com
 
     Args:
         job_id: Job ID to link summaries to
         host_count: Number of host summaries to create
+        job_created: Job creation timestamp to use for created date
+        job_finished: Job finished timestamp to use for modified date (aligns with real AWX behavior)
         unique_suffix: Optional unique suffix for host names (must match create_hosts suffix)
     """
     print(f'Creating {host_count} job host summaries for job {job_id}...')
@@ -491,7 +493,7 @@ def create_job_host_summaries(job_id, host_count, unique_suffix=None):
         failed = failures > 0 or dark > 0
 
         values.append(
-            f"(NOW(), NOW(), '{host_name}', {changed}, {dark}, {failures}, "
+            f"('{job_created}', '{job_finished}', '{host_name}', {changed}, {dark}, {failures}, "
             f'{ok}, {processed}, {skipped}, {str(failed).upper()}, NULL, '
             f'{job_id}, {ignored}, {rescued})'
         )