#4742·bullmq

[Bug]: Rust Worker never recovers from a dropped Redis connection

Author: willrnchCreated Sep 13, 2026Updated Sep 13, 2026
Labelsbug

Version

1.2.9

Platform

rust

What happened?

In the Rust client, every Redis connection (Worker, Queue, QueueEvents) is a bare redis::aio::MultiplexedConnection. Once the underlying socket dies (network blip, Redis restart), redis-rs never reconnects it on its own. Every subsequent command on that connection fails forever, and the worker silently stops fetching/processing jobs until the process is manually restarted. There's no automatic recovery path.

How to reproduce.

  • Create a worker
  • Push some job to the queue
  • Witness that the worker processes the jobs
  • Restart redis
  • Push more job
  • Witness that the worker doesn't process any job and doesn't throw any error

Code of Conduct

  • I agree to follow this project's Code of Conduct

Fix implemented here: https://github.com/willrnch/bullmq/commit/98b6b642b0ecbd75bf9381ffb2d574f57f148340

I'm not allowed to create a PR here.