Grafana MCP Server

Grafana MCP Server is the official Model Context Protocol server from Grafana Labs that connects AI assistants to your Grafana instance. With it, an LLM client such as Claude, ChatGPT or an IDE agent can search dashboards, query metrics and logs, dig into incidents and alerts, and summarize what is happening across your observability stack in natural language. The project is actively developed — it has around 3,600 stars on GitHub and receives updates in 2026.

Overview

The server runs over stdio or HTTP and exposes Grafana’s capabilities as MCP tools. It can search and retrieve dashboards, query Prometheus and Loki data sources with PromQL and LogQL, list and investigate Grafana Incident items, inspect alert rules and contact points, and pull data from data sources such as PostgreSQL. It works with Grafana open source, Grafana Cloud and Grafana Enterprise, authenticating with a service account token, and it supports read-only scoping so you control exactly what an agent is allowed to touch. For teams practicing SRE or on-call work, it turns routine triage — “why is this service erroring?” — into a conversation backed by live data.

Key Features

• Dashboard search, retrieval and summarization through any MCP-compatible AI client
• PromQL and LogQL queries against Prometheus and Loki data sources
• Grafana Incident investigation: list incidents, view activity and drill into timelines
• Alert rule and contact point inspection for faster alert triage
• Works with Grafana OSS, Cloud and Enterprise; service-account token authentication
• Fine-grained tool enable/disable controls and read-only operation modes

💳 Pricing

Grafana MCP Server is free and open source under the Apache-2.0 license. It connects to your existing Grafana deployment, whatever tier that runs on.

Who Is It For

It is for SREs, platform engineers and on-call teams who already run Grafana and want AI assistants to query their metrics, logs, incidents and alerts directly instead of copying screenshots into a chat window.