Improve ingestion concurrency safety and batch transaction robustness

This commit is contained in:
Alfredo Di Stasio
2026-03-10 16:37:29 +01:00
parent abd3419aac
commit 2252821daf
5 changed files with 202 additions and 44 deletions

View File

@ -1,8 +1,11 @@
import pytest
from contextlib import contextmanager
from celery.schedules import crontab
from django.utils import timezone
import psycopg
from django.conf import settings
from apps.ingestion.models import IngestionRun
from apps.ingestion.services.runs import _build_ingestion_lock_key, release_ingestion_lock, try_acquire_ingestion_lock
from apps.ingestion.tasks import scheduled_provider_sync, trigger_incremental_sync
from config.celery import app as celery_app, build_periodic_schedule
@ -44,23 +47,51 @@ def test_build_periodic_schedule_invalid_cron_disables_task_and_logs(settings, c
@pytest.mark.django_db
def test_trigger_incremental_sync_skips_overlapping_runs(settings):
def test_trigger_incremental_sync_skips_when_advisory_lock_not_acquired(settings, monkeypatch):
settings.INGESTION_PREVENT_OVERLAP = True
settings.INGESTION_OVERLAP_WINDOW_MINUTES = 180
IngestionRun.objects.create(
provider_namespace="mvp_demo",
job_type=IngestionRun.JobType.INCREMENTAL,
status=IngestionRun.RunStatus.RUNNING,
started_at=timezone.now(),
)
@contextmanager
def fake_lock(**kwargs):
yield False
monkeypatch.setattr("apps.ingestion.tasks.ingestion_advisory_lock", fake_lock)
run_id = trigger_incremental_sync.apply(
kwargs={"provider_namespace": "mvp_demo"},
).get()
skipped_run = IngestionRun.objects.get(id=run_id)
assert skipped_run.status == IngestionRun.RunStatus.CANCELED
assert "overlapping running job" in skipped_run.error_summary
assert "advisory lock" in skipped_run.error_summary
@pytest.mark.django_db
def test_advisory_lock_prevents_concurrent_acquisition():
provider_namespace = "mvp_demo"
job_type = IngestionRun.JobType.INCREMENTAL
lock_key = _build_ingestion_lock_key(provider_namespace=provider_namespace, job_type=job_type)
conninfo = (
f"dbname={settings.DATABASES['default']['NAME']} "
f"user={settings.DATABASES['default']['USER']} "
f"password={settings.DATABASES['default']['PASSWORD']} "
f"host={settings.DATABASES['default']['HOST']} "
f"port={settings.DATABASES['default']['PORT']}"
)
with psycopg.connect(conninfo) as external_conn:
with external_conn.cursor() as cursor:
cursor.execute("SELECT pg_advisory_lock(%s);", [lock_key])
acquired, _ = try_acquire_ingestion_lock(
provider_namespace=provider_namespace,
job_type=job_type,
)
assert acquired is False
cursor.execute("SELECT pg_advisory_unlock(%s);", [lock_key])
acquired, django_key = try_acquire_ingestion_lock(
provider_namespace=provider_namespace,
job_type=job_type,
)
assert acquired is True
release_ingestion_lock(lock_key=django_key)
@pytest.mark.django_db