Closed IceKhan13 closed 11 months ago
Probably it make sense to create a function wrapper for jobclient with retry logic since we are using it in multiple places.
Not sure how difficult they would be to write, but we should probably try adding tests for these types of scenarios
yes, I'm trying to carve out some time to boost test coverage for a project. It should not be that difficult. Hopefully in a week or so we have more time :)
retry logic since we are using it in multiple places
Yeah, I have the feeling that we are implementing a lot of retries in different places. Maybe something that we should start reviewing? Sounds relatively weird. Maybe it's normal until we can implement some kind of watchers or something.
What is the current behavior?
If cluster is not available for some reason job is failing in an update status loop.
What is the expected behavior?
client to connect to cluster should have retry logic to accommodate for cases when cluster is not yet available.