Comment by darkwater
1 day ago
Interesting project! Kudos for the release. One question: how are the failure scenario managed, i.e. AZP fails for whatever reason and it's in a crash loop? Just "no hints" to the scheduler, and that's it?
1 day ago
Interesting project! Kudos for the release. One question: how are the failure scenario managed, i.e. AZP fails for whatever reason and it's in a crash loop? Just "no hints" to the scheduler, and that's it?
If the AZP deployment fails, yes your correct there is no hints anywhere. If the lookup to AZP fails for whatever reason, it would be noted in the Kyverno logs. And based on if you -require- this policy to take affect or not, you have to decide if it you want pods to fail or not in the scheduling step. In most cases, you don't want to stop scheduling :)