What happened?
When the mongodb service is unable to respond to a request from the worker, the exception raised causes the worker to stop polling the request queue. I suspect in some cases the exception is caught somewhere outside of the main loop because the process does not stop so kubernetes does not restart the worker.
What are the steps to reproduce the bug?
Use a mongodb service that is somewhat unreliable.
Version
develop
Platform (OS and architecture)
Debian 12 x86_64
Relevant log output
No response
Accompanying data
No response
Organisation
MeteoSwiss