Production for Elasticsearchs
Running Elasticsearch in production means securing and observing the cluster. It includes enabling TLS and authentication, role-based access control, taking snapshots for backup, and watching metrics such as JVM heap, search latency, indexing rate and disk watermarks to catch problems early.
Stage: Operations · Part of the Elasticsearch roadmap · Last reviewed 19 November 2026
What to learn in Production
- Enable TLS and user authentication and create roles with index-level privileges
- Configure a snapshot repository and restore an index from a snapshot
- Monitor JVM heap usage, garbage collection and node disk space
- Explain what the disk watermark settings do when storage fills up
- Find slow queries with the slow log and the profile API
Production interview questions
What happens when a node's disk passes the flood-stage watermark?
Elasticsearch marks indices that have shards on that node as read-only, so writes are blocked to protect the node. Once you free disk space or add capacity, the block is released in recent versions or must be removed manually in older ones.
How do you back up an Elasticsearch cluster?
Register a snapshot repository such as shared storage or an object store, then take snapshots, ideally on a schedule with snapshot lifecycle management. Snapshots are incremental, and you restore indices from them. Copying data directories while running is not a safe backup.
Where Production fits in the Elasticsearch roadmap
Production belongs to the Operations stage (advanced level). Cluster health.
Other units in this stage
Back to the full Elasticsearch roadmap