Skip to main content
A queue worker processes dispatched jobs. The scheduler starts periodic maintenance, including recovery and cleanup. Run both services when your configuration uses both.

Match workers to enabled features

For example, if the configured connections are Redis and the assistant and archive queues are named media-studio-ai and media-archives, run a supervised process for each.
Set job, worker and retry timeouts for the corresponding feature. Use the archive guide for the longer archive job timeout. Ensure each worker can read and write its configured disks and run any processor binary it needs. Restart long-running workers after deploying updated code or configuration.

Verify scheduled tasks

The provider registers maintenance for navigation, workspace retirement and edited versions. Trash, Slack, API housekeeping, notifications and assistant maintenance depend on feature settings. The actual list for your installed app is authoritative. Add the schedules documented for upload receipts and staging and prepared archives if you enable those features. Coordinate scheduling across app replicas using shared cache locks or a single designated scheduler.

Test an interrupted workflow

In a test workspace, submit an image and an archive to their configured workers. Check that the uploaded original becomes usable independently of the thumbnail and that the ZIP becomes available only after its job finishes. Test recovery using a disposable transfer in the upload recovery guide. If an item stays pending, inspect the connection, queue name and failed jobs before retrying.