Open
Description
What is your proposal:
Coscheduling rely on delete from schedulingQueue, Coscheduling sibling Pods incurs overhead when in BackoffQ.
Why is this needed:
Is there a suggested solution, if so, please add it:
Kubernetes upstream has made this change. This issue was automatically resolved after Koordinator was upgraded to 1.32.