Search code examples
elasticsearchanomaly-detection

Anomaly detection for CPU usage

I have metric beat data from elasticsearch, which has a number of properties/columns like "cpu idle percentage", "cpu usage percentage" etc. I want to predict anomalies when my server cpu usage goes beyond 90% for lets say 3 minutes. Can someone shed some light on this?

Solution

  • Take a Look at this: Implementing a Statistical Anomaly Detector in Elasticsearch

    I've implemented the article myself and that's nice :-)