If you have a scenario like that in JENKINS-57396 the pod is never downloaded but the job just hangs and must be manually terminated.

It would be great if a timeout downloading a pod was considered a failure that could either be handled with a try {} in the pipeline script OR at minimum just failed the job.

This would allow for existing re-run functionality and better visibility from the UI that the job had failed rather than was still awaiting compute time.

I'm not familiar with the implementation details - it might be a sub-case of JENKINS-59340 but I felt like it was worth capturing independently.