gyfora commented on pull request #26: URL: https://github.com/apache/flink-kubernetes-operator/pull/26#issuecomment-1053918360
I agree this would be a good time to discuss and agree on a few design principles. In my view the Validate -> Observe -> Reconcile flow is very natural and easy to understand. We want to keep each of these components simple and self contained as much as possible. The commits that introduced jobmanager deployment validation (which is a very important, core feature) removed the original observer -> reconcile logic and moved part of the observing to the reconciler. To me this made the whole reconcile very complex and hard to understand with null checks and logic duplication. I think it leads to a simpler design to have a separation of concerns where the Observer component can describe the status of the deployment and the job, and the reconciler can act on it. If there is disagreement here we should probably move this to the mailing list for further discussion and visibility. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
