Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
None
-
None
-
None
Description
At the moment, computing the PreferredReplicaImbalanceCount metric traverses all the partitions in the cluster to find out the imbalance ones. This is extremely costly in cluster with large number of partitions and this is done after the processing of each event in the controller.
Attachments
Issue Links
- links to