I'm having an issue with CPU stats shown for Elasticsearch nodes in Kibana Monitoring.
The problem is that CPU Usage is always 0.
Kibana and Elasticsearch both run as docker containers and version 5.3.2.
Checking the kibana requests using the browser's developer tools I can see that the request to /api/monitoring/v1/clusters/uuid/elasticsearch/nodes returns:
"node_cgroup_quota": {
"metric": {
"app": "elasticsearch",
"field": "node_stats.process.cpu.percent",
"label": "Cgroup CPU Utilization",
"title": "CPU Utilization",
"description": "CPU Usage time compared to the CPU quota shown in percentage. If CPU quotas are not set, then the OS level CPU usage in percentage is shown.",
"units": "%",
"format": "0,0.[00]"
},
"min": null,
"max": null,
"slope": null
},
The weird thing is that if I go to a node's detail page it shows cpu usage:
Is there any configuration I'm missing? This was a 5.1.1 cluster that was upgraded to 5.3.2, and when it was 5.1.1 cpu stats were showing fine.
Thanks!