Mcp Grafana
Let your agent query Grafana dashboards, metrics, and related observability data from the stack you already run in production.
Overview
mcp-grafana is a MCP server for the Operate phase that exposes your Grafana instance dashboards and data to agents via stdio using URL and service-account or basic auth.
What is this MCP server?
- Official Grafana MCP server distributed as OCI image docker.io/grafana/mcp-grafana:0.9.0
- stdio transport with GRAFANA_URL required for your instance
- Auth via GRAFANA_SERVICE_ACCOUNT_TOKEN or GRAFANA_USERNAME and GRAFANA_PASSWORD
- GRAFANA_ORG_ID for multi-org setups or org header on HTTP transports
- Agent-facing path into existing Grafana dashboards and datasources
- Package version v0.9.0 as docker.io/grafana/mcp-grafana:0.9.0
- stdio transport with required GRAFANA_URL environment variable
- Optional GRAFANA_ORG_ID for multi-org Grafana deployments
Community signal: 3.1k GitHub stars.
What problem does it solve?
During incidents you waste time manually searching Grafana while your agent sits outside your observability system.
Who is it for?
Solo builders who already use Grafana and want Claude Code or Cursor to read the same panels and metrics during ops work.
Skip if: Greenfield projects with no Grafana yet, or teams that forbid agent access to production observability credentials.
What do I get? / Deliverables
After registering the Grafana MCP server with URL and credentials, your agent can pull dashboard and data context aligned with your live monitoring setup.
- Agent-driven queries against your Grafana dashboards and backing data
- Faster incident triage loops without leaving the MCP chat surface
Recommended MCP Servers
Journey fit
Grafana access matters when the product is deployed and you are interpreting live signals, not when you are only sketching UI mockups. Monitoring is the canonical shelf because the server description centers on dashboards and Grafana data access for ongoing ops.
How it compares
First-party Grafana observability MCP, not a Python sandbox or macro research feed.
Common Questions / FAQ
Who is mcp-grafana for?
Builders and small teams operating Grafana-backed stacks who want MCP agents to query dashboards and related data safely.
When should I use mcp-grafana?
Use it during operate and monitoring when you debug production issues, tune alerts, or explain metric spikes with agent help.
How do I add mcp-grafana to my agent?
Run the OCI image docker.io/grafana/mcp-grafana:0.9.0 over stdio, set GRAFANA_URL and GRAFANA_SERVICE_ACCOUNT_TOKEN (or username/password), and optional GRAFANA_ORG_ID in your MCP client config.