Grafana Integration

Export OpenClaw traces to Grafana Tempo and metrics to Grafana Mimir (or Prometheus) for visualization in Grafana dashboards.

Grafana Cloud (Direct Export)

1. Get Your OTLP Credentials

1. Go to grafana.com → Your stack → Connections → OpenTelemetry
2. Note the OTLP endpoint and generate an API token

2. Configure the Plugin

{
  "plugins": {
    "entries": {
      "otel-observability": {
        "enabled": true,
        "config": {
          "endpoint": "https://otlp-gateway-<region>.grafana.net/otlp",
          "protocol": "http",
          "serviceName": "openclaw-gateway",
          "headers": {
            "Authorization": "Basic <base64-of-instanceId:apiToken>"
          }
        }
      }
    }
  }
}

Creating the Basic auth header

echo -n "<instanceId>:<apiToken>" | base64

Self-Hosted Grafana Stack

Docker Compose Addition

Add Tempo and Grafana to the collector setup:

services:
  otel-collector:
    # ... existing config ...

  tempo:
    image: grafana/tempo:latest
    container_name: openclaw-tempo
    ports:
      - "3200:3200"   # Tempo API
      - "4417:4317"   # OTLP gRPC (Tempo)
    volumes:
      - ./collector/tempo-config.yaml:/etc/tempo/config.yaml:ro
    command: ["-config.file=/etc/tempo/config.yaml"]

  grafana:
    image: grafana/grafana:latest
    container_name: openclaw-grafana
    ports:
      - "3000:3000"
    environment:
      - GF_AUTH_ANONYMOUS_ENABLED=true
      - GF_AUTH_ANONYMOUS_ORG_ROLE=Admin
    volumes:
      - grafana-data:/var/lib/grafana

volumes:
  grafana-data:

Tempo Configuration

Create collector/tempo-config.yaml:

server:
  http_listen_port: 3200

distributor:
  receivers:
    otlp:
      protocols:
        grpc:
          endpoint: 0.0.0.0:4317

storage:
  trace:
    backend: local
    local:
      path: /tmp/tempo/blocks
    wal:
      path: /tmp/tempo/wal

metrics_generator:
  storage:
    path: /tmp/tempo/generator/wal

Collector Config (Fan-Out)

Update collector/otel-collector-config.yaml to export to both Dynatrace and Tempo:

exporters:
  otlphttp/dynatrace:
    endpoint: "${DYNATRACE_ENDPOINT}"
    headers:
      Authorization: "Api-Token ${DYNATRACE_API_TOKEN}"

  otlp/tempo:
    endpoint: "tempo:4317"
    tls:
      insecure: true

service:
  pipelines:
    traces:
      receivers: [otlp]
      processors: [batch]
      exporters: [otlphttp/dynatrace, otlp/tempo]

Grafana Data Source

1. Open Grafana at http://localhost:3000
2. Go to Configuration → Data Sources → Add data source
3. Select Tempo
4. Set URL: http://tempo:3200
5. Click Save & Test

Grafana Dashboards

Explore View

1. Go to Explore
2. Select the Tempo data source
3. Search by service name: openclaw-gateway
4. Or search by span name: openclaw.agent.turn or tool.*

Example Dashboard Panels

Token Usage Over Time (Prometheus/Mimir):

sum(rate(openclaw_llm_tokens_total[5m])) by (model)

LLM Latency P95 (Prometheus/Mimir):

histogram_quantile(0.95, rate(openclaw_llm_duration_bucket[5m]))

Tool Call Rate (Prometheus/Mimir):

sum(rate(openclaw_tool_calls_total[5m])) by (tool_name)

Error Rate (Prometheus/Mimir):

sum(rate(openclaw_llm_errors_total[5m])) / sum(rate(openclaw_llm_requests_total[5m])) * 100

Metric name format

OTel metrics use dots (openclaw.llm.tokens.total) but Prometheus converts them to underscores (openclaw_llm_tokens_total).