What is OpenSearch?
OpenSearch is an open-source, Elasticsearch-compatible search and analytics engine on TurboStack, with nightly index snapshots.
OpenSearch is an open-source search and analytics engine - a community-driven fork of Elasticsearch with a compatible API. It powers the same use cases (full-text product search, filtering, autocomplete, faceted navigation) and scales across nodes on the JVM.
Note
OpenSearch and Elasticsearch are alternatives - enable one per host.
On TurboStack
Enable OpenSearch at the host level and pick a version (opensearch_version, e.g. "2.x" or "3.x").
TurboStack auto-sizes the JVM heap; you can configure plugins. TurboStack takes nightly snapshots of
OpenSearch indices for backup and recovery. Choose OpenSearch when you want a fully open-source
engine.
Best practices
- Pick OpenSearch or Elasticsearch based on your application's compatibility.
- Keep the heap auto-sized unless you measure memory pressure - see Performance tuning.
- Rely on the nightly snapshots for index recovery; reindex after large data changes.