If some failures occur during any stages .e.g. scale up runner failed, the workflow job never gets to know. It keeps hanging for a new runner in case of ephemeral runner. How to handle such scenario. How to let the workflow job know that the job has failed.