Changed JVM heap usage pattern on 7.8.1

Elastic Stack Elasticsearch
1

Hi

Having recently upgraded to 7.8.1 on all our clusters , we are seeing changed JVM heap usage.

image
image918×261 35.3 KB

Our max heap usage seems to never go above around 19 GB , no matter what happens in the cluster. Same pattern on all nodes.

We upgraded from 7.6.2 but after upgrade we have started seeing long GC's again, which has not happened in a long time on 7.6.2. GC's in the 5-28 sec range , which is bad for us.

We are wondering with a 31 GB heap on 7.8.1 why it doesnt seem to utilize all available heap. It seems like quite a big part of the heap is reserved.

This is typical over 24 hours.

image
image915×257 35.5 KB

Any thoughts?

2

Can you post sections of your log showing the GC?

3

This topic was automatically closed 28 days after the last reply. New replies are no longer allowed.