Elasticsearch 6.3.2 master node high CPU & load average

I've upgraded from Elasticsearch version 2 to 6.3.2 on Debian Jessie from the official Elasticsearch repo.
After the upgrade I've started fresh and re-imported the data I needed.

I noticed that the cluster is sometime very slow with the master node busy in the 60-80% CPU and high load average. Index routing is super slow too after host leave or join the cluster.

The cluster has 7 hosts 3 master+data & 4 data only.
Each 2 CPU and 4GB RAM, all SSD.
It was working quite smoothly before the upgrade.

I am not sure where to start, would you have some advises please?

Thank you

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