Details
-
Improvement
-
Resolution: Fixed
-
Major
-
4.6.3
Description
As seen in one of the customer memory profiles, /getIndexStatus can generate huge amount of garbage as it initializes a new ClusterInfoCache every time it gets invoked and ns-server calls this every 1 second.
The memory profile is attached.
It would be good to cache the ClusterInfoCache results.
Attachments
Issue Links
- relates to
-
MB-29194 CLONE [MB-26671] - Optimize garbage generated by /getIndexStatus
- Closed
For Gerrit Dashboard: MB-26671 | ||||||
---|---|---|---|---|---|---|
# | Subject | Branch | Project | Status | CR | V |
87242,14 | MB-26671 - Optimize garbage generated by /getIndexStatus | unstable | indexing | Status: ABANDONED | +2 | 0 |
87357,2 | MB-26671 - Optimize garbage generated by /getIndexStatus | unstable | indexing | Status: ABANDONED | 0 | 0 |
87886,9 | MB-26671 - Optimize garbage generated by /getIndexStatus | unstable | indexing | Status: MERGED | +2 | +1 |
92689,2 | MB-26671 - Optimize garbage generated by /getIndexStatus | spock | indexing | Status: MERGED | +2 | +1 |