I have a windows service step that ocasionaly fails due to timeout on waiting for the windows service to start because the windows service in turn has API dependencies that take time to warm-up.
Normally the deployment succeeds if -retried immediatelly after failure and I decided to clone the step and run the clone only on failure of the previous (clone source) step.
This works as expected and the cloned step does kick-in and completes successfuly, however the release still fails because of the error in the original step!
My question: Is there a way to ignore the outcome of the original step if the step that kicks in on error succeeds?
I hope that makes sense - also open on suggestions how to best handle this timeout error in case my approach is wrong. I did think I could convert this Windows Service Install step to a script step and introduce a delay but it seemed like a hack and more work if it can be acheived more elegantly…