Details
-
Improvement
-
Status: Resolved
-
Major
-
Resolution: Fixed
-
3.0.0-alpha-1, 1.6.0, 2.5.0, 2.4.2
-
None
Description
In our production clusters, we have seen multiple cases where some of the wals are lingering in zk replication queues for months and we have no insight into it. Recently we fixed one case where wal was getting stuck if it is 0 size and from old sources. HBASE-25536. It would be helpful if we can have some metric which will tell us the age of oldest wal and we can add an alert for monitoring purpose.