grafana/mcp-grafana

What it solves

This project provides a Model Context Protocol (MCP) server that allows AI models (like Claude or Cursor) to interact directly with a Grafana instance. It bridges the gap between LLMs and observability data, enabling AI agents to search for dashboards, execute queries across various datasources, manage alerts, and handle incidents without manual intervention.

How it works

The server acts as a middleware that exposes Grafana's APIs to an MCP-compatible client. It provides a set of tools that the AI can call to perform actions in Grafana. To manage the context window of the LLM, it includes specific strategies like providing dashboard summaries instead of full JSON payloads.

Who it’s for

DevOps engineers, SREs, and platform engineers who use Grafana for monitoring and want to integrate AI agents into their observability workflows for automated troubleshooting, dashboard creation, and incident management.

Highlights

• Broad Datasource Support: Execute queries for Prometheus, Loki, InfluxDB, SQL (ClickHouse, Snowflake, Athena, etc.), CloudWatch, Graphite, Elasticsearch, and Quickwit.
• Dashboard Management: Search, create, update, and patch dashboards, with specialized tools to minimize LLM context window usage.
• Observability Operations: Manage Grafana Incidents, OnCall schedules, Alerting rules, and Sift investigations.
• Agent Observability: Specialized tools for Grafana Cloud users to inspect LLM conversations, evaluate scores, and manage test suites for AI agents.
• Infrastructure Integration: Support for generating accurate deeplinks to dashboards and panels, and rendering panels as PNG images.
• Configurable Toolset: Ability to disable specific tool categories to save context window space or restrict functionality.