Cluster Operations

Elasticsearch and OpenSearch clusters rarely fail from one big mistake — they fail from shards sized for last year’s data volume, ILM policies nobody revisited, and JVM heaps set once and never touched again. This hub collects the operational judgement calls: how to size and rebalance shards, how to build ILM policies that actually roll over and delete on schedule, how to plan a major-version upgrade without a multi-day outage, and how to read the cluster health signals before they turn into a red cluster status.

Reference

• Elasticsearch and OpenSearch version support matrix

Articles

• What ElasticDBA is

  A field guide to running Elasticsearch and OpenSearch in production — what this site covers and who it's for.

  2026-08-01