Configure SmithDB observability

Metrics

Each SmithDB component exposes application metrics at the fixed /metrics path on its HTTP port. Configure your monitoring stack to discover the pods by adding scrape annotations to each component:

Sample Prometheus configuration

smithdb:
  query:
    deployment:
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/path: "/metrics"
        prometheus.io/port: "8060"
  ingestion:
    deployment:
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/path: "/metrics"
        prometheus.io/port: "8050"
  compaction:
    deployment:
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/path: "/metrics"
        prometheus.io/port: "8070"
  compactionWorker:
    deployment:
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/path: "/metrics"
        prometheus.io/port: "9000"
  clusterManager:
    deployment:
      annotations:
        prometheus.io/scrape: "true"
        prometheus.io/path: "/metrics"
        prometheus.io/port: "8090"

These are Prometheus-style annotations. Providers with their own discovery format, such as Datadog, need that format applied to the same ports and path. Monitor the PostgreSQL metastore, Kubernetes nodes, and cloud services with your existing infrastructure monitoring.

For which metrics to watch and how to read them, see SmithDB metrics reference.

Logs and traces

SmithDB uses the chart-level OpenTelemetry tracing configuration described in Export LangSmith telemetry to your observability backend, with one difference: SmithDB exports over OTLP gRPC only. Set config.observability.tracing.exporter to grpc. With http, SmithDB export stays disabled and SmithDB logs remain console-only. The chart adds SmithDB pod and container names as OpenTelemetry resource attributes automatically.

To customize SmithDB log filtering, set RUST_LOG for all SmithDB workloads through smithdb.commonEnv:

smithdb:
  commonEnv:
    - name: RUST_LOG
      value: "warn,smithdb_query=debug,vortex=warn"

Confirm the collector is reachable from the LangSmith namespace, apply the chart, then confirm telemetry arrives. To set up or tune a collector, see Configure your collector for LangSmith telemetry.

Get started by cloud provider

Deploy with Terraform

Setup guides

Configuration

Connect external services

Platform auth & access control

SmithDB

Self-hosted observability

Hybrid

Scripts

Reference

Configure SmithDB observability

Metrics

Logs and traces

​Metrics

​Logs and traces

Metrics

Logs and traces