Discrepancy between CPU usage shown in Kibana vs AWS cloud watch metrics

I run Elasticsearch on EC2 instances launched in AWS. I use version 7.11.2. I've been load testing searches and I see CPU spikes up to 100% on data nodes when I look in Kibana. However, in AWS cloud watch I never see CPU spiked that much on any of the nodes, ever. What could possibly be the reason behind the discrepancy

How are you getting EC2 metrics into ES?

I'm usinng Kibana

Kibana offers several integrations to get your metrics in an Elasticsearch cluster. Are you using Metricbeat, as offered under standard AWS integration?

I'm not using metricbeat. I'm using the standard monitoring that comes out of the box.
Does it matter how kibana gets the values ? Whatever way, it should be same right ?

Your question is very vague, so in order for us to help you, I'm trying to get as much information as I can. Could be an ingestion problem, could be an aggregation problem, could be a visualization problem (unlikely, but still).

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