QueueScope

Articles & guides

Why do failed jobs disappear from a BullMQ dashboard?

A dashboard can inspect what Redis retains. If a failed job has been removed, changing the dashboard filter will not bring its payload or stack trace back.

Check your retention policy first

BullMQ supports automatic removal of completed and failed jobs. Producer options such as removeOnComplete and removeOnFail determine what remains available after processing. A count or age policy can bound retained jobs; removal may also be triggered by an explicit job action.

Review the options in the producer that added the job. A policy suited to a high-volume queue may intentionally keep fewer finished jobs than your debugging workflow needs. Retention is an application decision: keep the evidence you need while accounting for Redis memory and any sensitive payloads.

Confirm you are looking at the right scope

Before concluding that a job was deleted, check the Redis host, database, queue name, and BullMQ prefix. In QueueScope, confirm the connected profile rather than relying on edits in the connection form. Editing a saved profile does not change an active session.

If you know the job ID, use direct lookup. For broader discovery, inspect the correct job state and any active filters. Discovery and scans can be partial, so use Scan more or Search more when the interface says more data remains.

Read scan coverage alongside the count

Failure inbox scans retained failed jobs in bounded batches. The result reflects that scan’s coverage and observation times. A live queue can change between pages, so the scan is not a transactional snapshot. Switching the selected queue or all-queues scope resets scan state.

The inbox can compare newly observed jobs against the previous completed scan in the session. That comparison is not a guaranteed arrival rate or a permanent incident ledger. Jobs no longer retained cannot be reconstructed by QueueScope.

Decide what evidence should outlive the job

If you need a longer incident history, retain appropriate worker logs and application diagnostics outside the queue’s disposable job records. Avoid treating a completed/failed counter as a durable business audit trail. QueueScope’s local metric snapshots are useful for its local charts, but they are not a backup of deleted job payloads.

After investigating, make retention changes in your application configuration deliberately. Increasing retention may give future investigations more evidence; it will not restore data that was already removed.

Read the official BullMQ retention documentation and the QueueScope answer on missing jobs.