OpenSearch version 3.9 is now available on ClusterNest

OpenSearch 3.9 is live on ClusterNest, with a native neural sparse search engine, leaner vector storage, and new observability tools in Dashboards. The new neural sparse engine brings faster queries, faster index builds, and a significantly smaller heap footprint. For vector search, k-NN indexes can now store vectors in half-float format, and BF16 plus 2-bit and 4-bit scalar quantization give you more ways to trade memory for precision. Dynamic vector mapping lets you index vector data without defining the field mappings first. On the observability side, Dashboards adds a unified alerts view across data sources, a redesigned trace details page with a timeline waterfall, PromQL dashboard variables, and a no-code builder for Prometheus alerts. See the OpenSearch 3.9 announcement for the full rundown.

Provisioning and managing 3.9 clusters is supported today in the Console, the Terraform provider, and the MCP connector.

Try it out from the ClusterNest Console, update your Terraform configuration, or ask your AI agent to move existing clusters onto 3.9.