This PR supports resume the Flink statement with below different valid scenarios:
Resume under the original computePool and principalId.
Resume under a different computePool but original principalId.
Resume under a different principalId but original computePool.
Resume under a different computePool and different principalId.
And invalid scenarios:
Resume under an invalid computePool, the UPDATE request will fail immediately.
Resume under an invalid user account principalId, the UPDATE request will fail immediately.
Resume under an invalid service account principalId, the UPDATE request will pass but the statement itself will fail subsequently, and the failed statement can be resumed later.
New Features
The Flink statement resume support can use both the resume and update interfaces, which are functionally equivalent, for example:
`Request to resume statement "my-statement" with service account "sa-123456" and move to compute pool "lfcp-123456".`
confluent flink statement update my-statement --stopped=false --principal sa-123456 --compute-pool lfcp-123456
confluent flink statement resume my-statement --principal sa-123456 --compute-pool lfcp-123456
Checklist
[x] Leave this box unchecked if features are not yet available in production
:tada: All Contributor License Agreements have been signed. Ready to merge. Please push an empty commit if you would like to re-run the checks to verify CLA status for all contributors.
Release Notes
This PR supports resume the Flink statement with below different valid scenarios:
computePool
andprincipalId
.computePool
but originalprincipalId
.principalId
but originalcomputePool
.computePool
and differentprincipalId
.And invalid scenarios:
New Features
resume
andupdate
interfaces, which are functionally equivalent, for example:Checklist
What
References
Test & Review
Manual verification: https://docs.google.com/document/d/1q93WL5HXyh9i4143tQwbTyBs5LwIUbieiRL4IQT-PWc/edit?tab=t.0#heading=h.q4ezisb3sd4t