Container termination times out on Kubernetes

[JoshuaWatt]

When running zuul-weeder in Kubernetes, I noticed that the pod termination takes a long time (30 seconds). I suspect this is because zuul-weeder doesn't exit cleanly on SIGTERM and when the grace period elapses (30 seconds by default), it gets sent SIGKILL

[TristanCacqueray]

I also noticed a delay when using systemctl stop. Though it seems like SIGTERM is properly handled (at least when running from the command line), and the service stops with exit code 143, as expected. We'll have to check if there is something special to do for the container runtime.

Thank you for the bug report!