Cluster for Elasticsearchs
An Elasticsearch cluster is a group of nodes that together hold indices split into primary and replica shards. Cluster work covers node roles, shard allocation, how many shards to create, and sizing hardware so that search and indexing stay fast as data grows.
Stage: Operations · Part of the Elasticsearch roadmap · Last reviewed 19 November 2026
What to learn in Cluster
- Explain primary and replica shards and how replicas give both failover and read capacity
- Read cluster health colours and find unassigned shards with the allocation explain API
- Choose a shard count and size that avoids too many tiny shards
- Assign node roles such as master, data and ingest
- Use index lifecycle management to roll over and age out time-based indices
Cluster interview questions
What does a yellow cluster status mean?
All primary shards are allocated, so data is searchable, but one or more replica shards are not. This is normal on a single-node cluster, since a replica cannot sit on the same node as its primary. Red means at least one primary is unassigned.
Why are too many small shards a problem?
Every shard is a Lucene index with its own memory, file handles and cluster state overhead. Thousands of tiny shards waste heap and slow cluster operations. It is better to use fewer, larger shards and roll over time-based indices by size or age.
Where Cluster fits in the Elasticsearch roadmap
Cluster belongs to the Operations stage (advanced level). Cluster health.
Other units in this stage
Back to the full Elasticsearch roadmap