Grafana MCP Analyzer

What it does

Connects AI assistants to Grafana monitoring data, enabling them to read metrics and logs directly. It transforms raw monitoring charts from Prometheus, MySQL, and Elasticsearch into expert-level DevOps analysis, bridging the gap between raw data and operational insights.

Tools

• analyze_query: Fetches monitoring data and provides an initial AI analysis.
• analyze_existing_data: Performs deep-dive analysis based on previously fetched data in the session.
• chunk_workflow: Automatically handles large datasets by splitting them into manageable chunks for the LLM.
• manage_cache: Controls the session-based data cache to optimize performance and accuracy.
• list_queries: Lists all available data source queries configured in the system.
• check_health: Verifies the connection status and health of the Grafana server.
• list_data: Retrieves a list of historical data stored in the local cache.
• server_status: Provides operational details about the MCP server instance.

Installation

Add the following to your claude_desktop_config.json:

{
  "mcpServers": {
    "grafana": {
      "command": "grafana-mcp-analyzer",
      "env": {
        "CONFIG_PATH": "https://raw.githubusercontent.com/SailingCoder/grafana-mcp-analyzer/main/config/grafana-config-play.js",
        "MAX_CHUNK_SIZE": "100"
      }
    }
  }
}

Supported hosts

• claude
• cursor
• chatgpt