Metricbeat/docker shows different stats for logstash in container than logstash monitoring

Elastic Observability Metrics
1

I’m getting weird perf stats on logstash-netflow. Seems like the docker API perf mon says it is using 1200% CPU, but the actual logstash monitoring shows 8% CPU. Also htop also show lots of CPU headroom.

28
Screenshot 2019-07-24 at 12.39.28.png1770×894 275 KB
41
Screenshot 2019-07-24 at 12.36.41.png1762×1042 81 KB
34
Screenshot 2019-07-24 at 12.35.34.png2744×2314 263 KB

2

Hi @sgreszcz,

it's hard to determine the possible causes based on these screenshots. Here's a few things that might help narrow it down:

  • Which version of metricbeat and Kibana are you running?
  • What are the filter and group parameters used in the last screenshot?
  • Is the logstash process the only one running in the docker container?
  • Does docker stats report correct values for the container?
3

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