izaitsevfb/grafana-mcp
If you are the rightful owner of grafana-mcp and would like to certify it and/or have it hosted online, please leave a comment on the right or send an email to henry@mcphub.com.
This project provides an MCP server for Grafana, including tools to interact with Grafana dashboards, data sources, alerts, and more.
Grafana MCP
This project provides MCP server for Grafana, including MCP tools to interact with Grafana dashboards, data sources, alerts, and more.
The package includes a dashboard template resource that is properly bundled with the package for use when installed.
Usage
Installation
- Create a virtual environment (recommended)
python -m venv venv
source venv/bin/activate
- Install this package directly from GitHub using pip:
pip install -e git+https://github.com/izaitsevfb/grafana-mcp.git#egg=grafana_mcp
(Alternatively, you can install it locally with pip install -e .)
- Configure Grafana authentication
Create a .env file with your Grafana URL and API token:
cp .env.example .env
Then edit the .env file to add your actual Grafana URL and API token:
GRAFANA_URL=http://your-grafana-server:3000
GRAFANA_API_TOKEN=your-api-token-here
You can generate an API token in Grafana by navigating to: Configuration → API Keys → New API key.
- After installation, the MCP server can be run as:
python -m grafana_mcp
- Add this MCP server to Claude Code:
claude mcp add-json grafana '{ "type": "stdio", "command": "python", "args": [ "-m", "grafana_mcp" ], "env": {} }'
Note: By default, MCP config applies only to the current directory. If you want to use it globally, add --scope user to the command above:
claude mcp add-json --scope user grafana '{ "type": "stdio", "command": "python", "args": [ "-m", "grafana_mcp" ], "env": {} }'
- Run Claude Code as usual:
claude
Development
Development Installation
- Clone the repository
- Install in development mode:
pip install -e .
The package uses Python's package resource management system to include the dashboard template. The dashboard.json file is stored in the src/grafana_mcp/ directory and is accessed using importlib.resources when the package is installed.
Testing
Run unit tests:
python -m unittest discover tests
For testing with token authentication, ensure you have created a .env file with your Grafana credentials as described above. The test suite includes mocked tests that don't require an actual Grafana instance.
Tools
The Grafana MCP provides the following tools:
Available Tools
Dashboard Management
get_grafana_info(): Get information about the connected Grafana instance including version and connection statuslist_dashboards(): List all dashboards from the connected Grafana instanceget_dashboard(uid): Get details about a specific dashboard by UID
Organization Management
get_organization(): Get information about the current organization
Data Source Management
list_datasources(): List all available data sources
Requirements
grafana-client: Python client for Grafana APIpython-dotenv: For loading environment variables from.envfilemcp: Model Context Protocol server implementation
Authentication
This package uses token authentication with the Grafana API. To configure authentication:
- In your Grafana instance, create an API key with appropriate permissions
- Create a
.envfile with the following variables:GRAFANA_URL=http://your-grafana-url:3000 GRAFANA_API_TOKEN=your-api-token-here - The MCP server will automatically load these credentials when started
More tools and features will be added in future versions.
