Details
-
Task
-
Resolution: Obsolete
-
Undefined
-
None
-
None
-
False
-
None
-
False
-
0
Description
We periodically experience full PVC volume alerts from different indexers in production. Because the alerts happen ~1 week after continuous traffic there is likely some corner-case that leads to the indexer not cleaning up layers after indexing (maybe in a failure scenario?).
Example alert: https://coreos.slack.com/archives/C5QAR41JM/p1651329788707589
Volume size: 100Gb