How do you handle poison messages that crash workers?
PICTURE THIS: STACK VS QUEUE
Simple meaning
Cap retries, catch expected exceptions, and route failures to a dead-letter queue or an error table.
WHY — Celery instead of guessing?
Why interviewers care about Celery:
question about Celery.
trade-offs, and what you would actually do on a Python project - not buzzwords.
Name the idea, why it exists, then one short example.
End with when you use it and one common pitfall.
STEPS — What happens step by step?
Before you speak the answer, walk the interviewer through these steps:
- 1Cap retries, catch expected
exceptions, and route failures to a dead-letter queue or an error table.
- 2Reject + requeue forever
will take down the cluster.
- 3Visibility into task_id, args
(redacted), and traceback belongs in Sentry or structured logs, not in the broker payload forever.
- 4Give an example
One tiny concrete case you can say aloud.
- 5Common mistake
What juniors usually get wrong.
- 6Close
When you pick this over the alternative.
EXAMPLE — See it in action
Here's a short line you can speak, broken into clear beats:
Note: Adapt this scaffold to your own project — keep it under 60–90 seconds.
Key takeaway
Cap retries, catch expected exceptions, and route failures to a dead-letter queue or an error table. Reject + requeue forever will take down the cluster.