Operations Playbook: Celery Backlog¶

Trigger Conditions¶

• queue depth grows continuously
• async outcomes delayed beyond SLA
• repeated retries/failures in worker logs

Response Workflow¶

1. Confirm worker liveness.
2. Inspect registered tasks and ping.
3. Validate Redis health/connectivity.
4. Identify failing task class patterns.
5. Restart/scale worker if needed.
6. Validate backlog drain trend.

docker compose logs -f celery redis
docker compose run --rm celery celery -A app inspect ping
docker compose run --rm celery celery -A app inspect registered

Recovery Validation¶

• backlog trend stabilizes or declines
• failed/retried task rates improve
• user-visible async outcomes recover