Skip to main content
WednesdayAI exports observability data via OpenTelemetry (OTel). Route this to Langfuse, Grafana Tempo, Jaeger, or any OTel-compatible backend.

​
Quick start with Langfuse

1

Create a Langfuse project

Sign up at langfuse.com and create a project. Note the public key, secret key, and host URL.
2

Configure the OTel exporter

diagnostics:
  otel:
    enabled: true
    endpoint: "https://cloud.langfuse.com/api/public/otel/v1/traces"
    headers:
      Authorization: "Basic <base64(public_key:secret_key)>"
3

Verify

openclaw restart && openclaw doctor --check otel

​
Quick start with Grafana / Tempo

diagnostics:
  otel:
    enabled: true
    endpoint: "http://your-grafana-agent:4318/v1/traces"
    headers: {}

​
What is traced

SpanDescription
gateway.requestTop-level span for each incoming message
llm.completeLLM API call — model, token counts, latency
hook.dispatchEach hook invocation with name and duration
tool.callAgent tool calls
channel.sendOutbound message delivery

​
Prometheus metrics

When diagnostics.enabled: true, a /metrics endpoint is exposed in Prometheus format.
MetricDescription
wednesdayai_requests_totalTotal requests by channel and status
wednesdayai_llm_tokens_totalLLM tokens consumed by model
wednesdayai_llm_latency_msLLM request latency histogram
wednesdayai_active_sessionsCurrently active sessions