Connect OpenClaw to Grafana
Let an OpenClaw agent read Grafana dashboard data, check the health of your Grafana, Loki and Mimir stack, and ship logs to Loki, with no credentials handed to the agent.
Grafana is the dashboard layer most engineering teams put in front of their metrics, logs and traces. Together with Loki for logs and Mimir for Prometheus metrics, it is where on-call engineers go to answer “is it broken, and since when?”. The dashboards are excellent, but somebody still has to open them. Reading the same panel each morning, pasting a screenshot into a status thread, and checking whether the ingestion ring is healthy after a deploy are all small jobs that add up.
With Claworc’s External Connections, an OpenClaw agent can talk to
your Grafana stack on your behalf. You connect Grafana once through the instance
settings, Composio holds the resulting token, and the agent gets an auto-generated
claworc-grafana skill that tells it which Grafana tools exist and how to call them.
The agent never sees a password, a bearer token or an API key.
Tools available to OpenClaw agents
Section titled “Tools available to OpenClaw agents”Composio’s Grafana toolkit exposes 11 tools. Several of them target the operational endpoints of Loki and Mimir rather than the dashboard UI, which makes the toolkit particularly useful for health checks. In practice an OpenClaw agent leans on a handful of capabilities:
- Read dashboard data — query a panel on a public Grafana dashboard and get the time-series data back, which lets the agent read a chart instead of looking at it.
- Check Grafana health — confirm the Grafana instance is up and its database is reachable, and whether a valid Grafana Enterprise license is in place.
- Ship logs to Loki — push OpenTelemetry logs into Loki, so the agent can write its own activity into the same place your application logs already live.
- Inspect the HA tracker — see which replica Mimir has elected as leader for each Prometheus HA pair.
- Inspect hash rings — check the rings for Mimir’s ruler and Loki’s index gateway to confirm every instance is active.
- List store-gateway tenants — see which tenants have data on a store-gateway node.
Every action goes through a small broker running next to the agent, so the agent only needs to describe what it wants done in plain language and let the skill handle the rest.
Use cases
Section titled “Use cases”A dashboard readout in words
Section titled “A dashboard readout in words”“What does the checkout latency panel look like for the last six hours?” The agent pulls the series from the dashboard and answers with numbers: the p95 now, the peak, when it started climbing. Ask for it every morning and the daily status thread writes itself.
Post-deploy health check
Section titled “Post-deploy health check”After rolling out a new Loki or Mimir version, ask the agent to confirm the deployment is healthy. It checks Grafana’s health first, then the ruler and index gateway rings and the HA tracker, and reports any instance that is not in the expected state before the next alert does.
Keep agent activity in Loki
Section titled “Keep agent activity in Loki”Have the agent send a short OTLP log line to Loki each time it completes a task in another connected service. Its actions then show up next to your application logs, searchable with the same LogQL you already use.
Get started
Section titled “Get started”Connecting Grafana takes three steps: an admin adds a Composio API key to Claworc’s
settings, you open the instance’s Settings tab, click + Add connection, pick
Grafana and authorize the connection. The claworc-grafana skill appears in the instance
a moment later. The full walkthrough, including the exact Composio permissions the key
needs, is in the External Connections documentation.