During graceful shutdown, AM's TimerPool logs an error for each thread which it needs to interrupt to prompt its shutdown. Unfortunately, the error message doesn't log the threads name (only its thread ID) so it's difficult to establish whether or not this truly is a problem.
This issue is seen when running AM in Kubernetes on GCP as part of the saas.
- Start one or more AM pods in Kubernetes.
- Tail the logs from an AM pod using kubectl logs -f <am pod name> -c openam
- Delete AM pod using GCP console