Skip to main content
SmithDB is an optional, opt-in datastore for self-hosted LangSmith. It is available in beta. ClickHouse remains required. You cannot remove ClickHouse after enabling SmithDB. Minimum LangSmith version depends on cloud. See Check cloud support.
SmithDB can serve LangSmith trace ingestion and queries alongside the existing ClickHouse datastore. Enable it on a Kubernetes installation when you want SmithDB-backed query performance while keeping ClickHouse in place. This section is for existing self-hosted installations. Complete the Kubernetes setup guide first. Do not treat SmithDB as part of a new required install path.

​
Check cloud support

SmithDB examples in this section cover managed Kubernetes on AWS, GCP, and Azure. They do not cover self-managed clusters.

Install SmithDB

Stage infrastructure, deploy SmithDB services, enable dual ingestion, and switch queries after validation.

Prepare supporting infrastructure

Provide a dedicated PostgreSQL metastore, object storage, and local SSD cache capacity.

Configure for scale

Choose a tested resource tier and size SmithDB workloads.

Configure observability

Scrape SmithDB metrics and export logs and traces to your stack.

Migrate historical data

Backfill ClickHouse history into SmithDB before query cutover.

Troubleshoot SmithDB

Gather context, disable SmithDB, or reset a deployment safely.

​
Keep required storage

A self-hosted LangSmith instance still uses ClickHouse for traces and feedback, PostgreSQL for operational data, Redis or Valkey for queuing and caching, and optional blob storage. SmithDB adds its own metastore, object store, and cache nodes. It does not replace those required services in this release. For SDK method changes that work with both ClickHouse and SmithDB, see Migrate to SmithDB-backed SDK methods.

​
See also

Connect these docs to Claude, VSCode, and more via MCP for real-time answers.
Edit this page on GitHub or file an issue.