Nodes with too long GC pauses with G1GC

[mikkolehtisalo]

Graylog generates a lot of GC pause alerts like

Nodes with too long GC pauses (triggered 2 days ago) There are Graylog nodes on which the garbage collector runs too long. Garbage collection runs should be as short as possible. Please check whether those nodes are healthy. (Node: 7ee19339-0bae-45fe-9325-56166bcf088e, GC duration: 1699 ms, GC threshold: 1000 ms)

after switching to G1 GC. I believe the alerts are in error. GC probably does take a long time with G1GC, but as the world isn't stopped it isn't an issue. This alert should probably be modernized for use with newer JVMs with newer garbage collectors.

[zino7825]

I've same experienced after upgrade java from 8 to 11.

[mikkolehtisalo]

For now it is possible set gc_warning_threshold = 30s as workaround.

[boosty]

https://github.com/Graylog2/graylog2-server/pull/13929#issuecomment-1326126342