Handle Failed Jobs¶

Understand Failure Paths¶

In worker runtime:

• retries available -> job becomes delayed (backoff) and is re-scheduled
• retries exhausted -> job becomes failed and is moved to DLQ

Inspect Failures¶

asyncmq job list --queue default --state failed
asyncmq queue info default

Dashboard also exposes DLQ actions under /queues/<name>/dlq.

Retry Failed Jobs¶

From CLI:

asyncmq job retry <job_id> --queue default

From dashboard:

• open queue DLQ page
• select job(s)
• run retry action

Remove Poison Jobs¶

asyncmq job remove <job_id> --queue default

Use removal only when retry is known to be unsafe or irrelevant.

Prevent Repeat Incidents¶

• make handlers idempotent
• cap retries and use backoff
• keep payloads small and deterministic
• add handler-level input validation