Grafana

Grafana MCP Connector for Claude

A+

Manage observability via Grafana — search and inspect dashboards, monitor alerts, and handle data sources directly from any AI agent.

4 tools Official Updated Jun 28, 2026 Official Vinkius Partner

Connect your Grafana instance to any AI agent and take full control of your application observability, dashboard inspections, and alerting workflows through natural conversation.

What you can do

  • Dashboard Discovery Orchestration — Search for Grafana dashboards by title or tag to retrieve unique UIDs and metadata natively within your chat
  • Panel & Query Inspection — Retrieve the full configuration of any dashboard by UID, extracting precise PromQL, LogQL, or SQL queries and panel layouts flawlessly
  • Data Source Auditing — List all configured data sources including Prometheus, Loki, CloudWatch, and SQL databases to verify connectivity boundaries securely
  • Alert Monitoring Oversight — Enumerate active alert rules and retrieve current firing states to monitor system health and resolve incidents synchronously
  • Observability Navigation — Analyze specific localized variables decoding active monitoring routes and extracting structural constraints from your Grafana environment
  • SRE Command Center — Verify dashboard UIDs and retrieve query strings to debug performance regressions or analyze log patterns using natural language

How it works

  1. Subscribe to this server
  2. Enter your Grafana Instance URL and Service Account Token (found in Administration > Users & Access > Service accounts)
  3. Start inspecting your observability stack from Claude, Cursor, or any MCP-compatible client

Who is this for?

  • SRE & DevOps Engineers — monitor alert rules and inspect dashboard queries without manual navigation in the Grafana UI
  • Cloud Architects — audit data source configurations and verify dashboard organization using natural language
  • Software Developers — extract PromQL or LogQL queries from existing panels to verify metrics and logs during development
  • Incident Responders — quickly find relevant dashboards and check alert statuses during active troubleshooting
dashboardsmetricslogstracesalertingdata-visualization

4 tools expose this connector's capabilities to your AI agent.

search_dashboards

Returns basic info including the UID. To inspect the panels and queries of a dashboard, use get_dashboard with the uid. Search Grafana dashboards by title or tag

get_dashboard

Requires the dashboard UID, which you can get from search_dashboards. Get full dashboard configuration including panels and queries

list_datasources

) are available in this Grafana instance. List all configured data sources in Grafana

firing_alerts

Returns alerting rules that are currently in "firing" state, including their labels and annotations. Get currently firing alerts from Grafana Unified Alerting

See how to talk to your AI agent using Grafana.

Search for dashboards tagged with 'production'

Searching dashboards... I found 3 matching dashboards: 'K8s Cluster Overview' (UID: abc-123), 'HTTP Response Metrics' (UID: def-456), and 'Database Performance'. Which one would you like me to inspect for panels?

Show me the queries for dashboard 'abc-123'

Retrieving dashboard 'abc-123'... I found 5 panels. Highlights: 'CPU Usage' uses PromQL `sum(rate(node_cpu_seconds_total...))`, and 'Request Latency' uses `histogram_quantile(0.95, sum(rate(http_request_duration_seconds_bucket...)))`. Would you like the full JSON or more queries?

Are there any firing alerts right now?

Checking alert rules... Yes, I found 1 firing alert: 'High Error Rate' on service 'API-v1'. It has been firing for 15 minutes. All other 12 alert rules are currently in 'Normal' state.

Yes. Use the 'search_dashboards' tool. You can provide an optional query string to match titles or tags. The agent will return basic info including the unique UID required for deeper inspection.

Related Connectors