Instrumenting CPU usage of logstash and elasticsearch

We made some changes to our elasticsearch install (upgrading to 6, enabling beats, pipeline changes...) and the CPU usage of our logstash/elastic server has shot up:

The CPU% usage is roughly split between logstash and elasticsearch.

How can we instrument the stack to determine WHY it's using so much CPU (ingest rate of approximately 5K/s) and change how we're doing things (if possible) to use less CPU?

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