Lakshmi

@davidbirdsong Thanks for posting this idea! I think this would be a cool feature. We support some part of this behavior in the [Recovering](https://github.com/lyft/flinkk8soperator/blob/master/pkg/controller/flinkapplication/flink_state_machine.go#L440) state, except that in the current...

Hi @nealsun, Thanks for the question. In order to migrate a Flink job to run on `flinkk8soperator`, you can update the SavepointInfo section of the CRD with a SavepointLocation. Code:...

We've spec-ed out a plan for supporting zero-downtime deploys. The design document can be found here: https://docs.google.com/document/d/14OreDj1I-EtlvEPpXN2KWPMo5zlVJRm_rdGQ00ZFBxg/edit#heading=h.h5422sqgwdpj Please let us know your feedback/comments!

Hi @davidbirdsong ! Thank you for creating this PR. I have one overall comment/suggestion on the approach. Including the functionality you have in this PR, we have 3 ways to...

Sorry for the delay here @davidbirdsong ! Thanks for addressing the review comments. I am leaning on the approach where we introduce an `UpdateMode`, which can default to the savepoint...

Hey @davidbirdsong ! Thanks for the updates. I will review this PR in the next couple of days :)

That's a very good point! I don't believe we'd want the `updateMode` and `savepointPath` to interact at all in fact, because the expectation today is that the user-provided `savepointPath` is...