Draft: Make Kubernetes executor retry error configurable
What does this MR do?
Job run with the executorkubernetes can fail for multiple reasons. Some of the errors returned after failure can/should trigger a retry of the failed k8s API calls. This MR allows user to configure error that can/should trigger a retry
Why was this MR needed?
To give more flexibility on error that can trigger a retry of a k8s API call failure
What's the best way to test this MR?
TDB