Analyzing slow queries and improving their performance can be very challenging. This section examines how to look for the root cause of poor query performance, and it offers some different approaches to finding a solution.
If you notice poor query performance, start with the slow log. To enable the slow log, edit
elasticsearch.yml and add these configuration options to all nodes on the cluster:
index.search.slowlog.threshold.query.warn: 8s index.search.slowlog.threshold.query.info: 4s index.search.slowlog.threshold.query.debug: 2s index.search.slowlog.threshold.query.trace: 500ms index.search.slowlog.threshold.fetch.warn: 1s index.search.slowlog.threshold.fetch.info: 750ms index.search.slowlog.threshold.fetch.debug: 500ms ...