Issue Summary

This is reproducible on Data Center: (yes) 

Steps to Reproduce

1. Enable ephemeral agent support 
2. Create a Bamboo build plan to run on an Ephemeral agent by Dedicating ephemeral agent templates.
3. Run the build plan and wait for the Ephemeral agent's dedicated Kubernetes pod to get into running state.
4. Delete the Ephemeral agent's dedicated Kubernetes pod or shutdown the Kubernetes worker node hosting the Ephemeral agent's dedicated Kubernetes pod.

Expected Results

Bamboo Data Center should recreate the Ephemeral agent's dedicated Kubernetes pod when the pod gets destroyed.

Actual Results

The below messages are seen in the  <BAMBOO-HOME>/logs/atlassian-bamboo.log file:

2024-08-19 21:37:46,222 INFO [AtlassianEvent::1-BAM::EVENTS:pool-2-thread-45] [BuildQueueManagerImpl] Job expects an ephemeral agent, skip broadcasting.
2024-08-19 21:37:47,372 INFO [AtlassianEvent::1-BAM::EVENTS:pool-2-thread-48] [KubernetesExecutor] Launching Ephemeral Agent pod chec-bas-job1-12-ovxlgsrm for checkpoint - basic - Default Job #12 (CHEC-BAS-JOB1-12

...
2024-08-19 21:43:04,179 WARN [AtlassianEvent::1-BAM::EVENTS:pool-2-thread-12] [BuildQueueTimeoutNotificationListener] Build queue timeout (5 minutes) has been exceeded for 'CHEC-BAS-JOB1-12' 

Bamboo Data Center continues to keep the build on a queue without acknowledging the Ephemeral agent's dedicated Kubernetes pod that is destroyed.

Workaround

Stop the build and Manually run the build again.