IOPS change monitoring



I have a 2.4 cluster running on AWS instances.
I want to lower the disk's IOPS and would like to monitor the affect on the performance of ES.

  1. Which operations will be affected from IOPS decrease?
  2. How can I monitor the performance decrease (Which end point, which metric)?
  3. How will I know the IOPS is to low and the cluster is no longer working good?


(Nik Everett) #2

Indexing always uses the disk. Searching will use the disk if it the files aren't already in the filesystem cache.

The best metric to monitor is metrics from your application like search response time. You should also look for Elasticsearch logging stuff about merging falling behind.

(system) #3

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