glances/mcp

Glances MCP Server

A Model Context Protocol (MCP) server that provides system monitoring capabilities through integration with Glances, a cross-platform system monitoring tool.

Overview

The Glances MCP Server exposes Glances monitoring data through the MCP protocol, allowing AI assistants and other MCP clients to query system statistics, monitor processes, check resource usage, and analyze system performance.

Features

• Real-time System Monitoring: Access CPU, memory, disk, network, and process information
• Multiple Transport Options: Supports both stdio (standard) and HTTP transports
• 15+ Tools: Comprehensive set of tools for system monitoring
• Resource Summaries: Formatted system statistics summaries
• Performance Analysis: Built-in prompts for system performance analysis

Prerequisites

• Python 3.10 or higher
• Glances installed and running
• pip or uv package manager

Installation

1. Install Glances

# Using pip
pip install glances

# Using system package manager (Ubuntu/Debian)
sudo apt install glances

# Using Homebrew (macOS)
brew install glances

2. Install MCP Server Dependencies

# Install required packages
pip install requests mcp uvicorn fastapi

Or create a requirements.txt:

requests>=2.31.0
mcp>=1.0.0
uvicorn>=0.27.0
fastapi>=0.104.0

Then install:

pip install -r requirements.txt

3. Download the MCP Server

Save the glances_mcp.py file to your desired location.

Usage

Starting Glances

First, start Glances in web server mode:

glances -w

By default, Glances runs on http://localhost:61208

Running the MCP Server

Option 1: Stdio Transport (Default - for Claude Desktop)

python glances_mcp.py

This mode is used for direct integration with Claude Desktop and other MCP clients that communicate via stdio.

Option 2: HTTP Transport (for testing and remote access)

# Using uvicorn
uvicorn glances_mcp:app --host 0.0.0.0 --port 8000

# Or with auto-reload for development
uvicorn glances_mcp:app --host 0.0.0.0 --port 8000 --reload

# Or using the built-in flag
python glances_mcp.py --http

The MCP endpoint will be available at: http://localhost:8000/mcp

Configuration

Glances URL

By default, the server connects to Glances at http://localhost:61208. To change this, edit the GLANCES_URL variable in glances_mcp.py:

GLANCES_URL = "http://your-glances-server:61208"

Claude Desktop Integration

Add the following to your Claude Desktop configuration file:

macOS: ~/Library/Application Support/Claude/claude_desktop_config.json Windows: %APPDATA%\Claude\claude_desktop_config.json Linux: ~/.config/Claude/claude_desktop_config.json

For Stdio Transport:

{
  "mcpServers": {
    "glances": {
      "command": "python",
      "args": ["/full/path/to/glances_mcp.py"]
    }
  }
}

For HTTP Transport:

{
  "mcpServers": {
    "glances": {
      "url": "http://localhost:8000/mcp"
    }
  }
}

Available Tools

The MCP server provides the following tools:

System Status

• get_status: Check Glances API health status
• get_all_stats: Get all system statistics from all plugins
• get_plugins_list: Get list of available Glances plugins

CPU & Memory

• get_cpu_stats: Get CPU usage statistics
• get_memory_stats: Get memory usage statistics

Storage & I/O

• get_disk_stats: Get disk I/O statistics
• get_sensors: Get sensor readings (temperature, fans)

Network

• get_network_stats: Get network interface statistics

Processes

• get_processes: Get process list (optionally limit to top N processes)
• get_process_by_pid: Get detailed information for a specific process

System Information

• get_system_info: Get system information (OS, hostname, platform)
• get_config: Get Glances configuration

Docker

• get_docker_stats: Get Docker container statistics (if Docker is available)

Advanced

• get_plugin_data: Get data from any specific Glances plugin by name
• get_plugin_history: Get historical data for a plugin
• get_plugin_limits: Get configured limits for a plugin

Resources

• glances://stats: A formatted summary of system statistics including CPU, memory, and load averages

Prompts

• analyze_system_performance: A template prompt for comprehensive system performance analysis

Testing

Using MCP Inspector

The MCP Inspector is a great tool for testing your MCP server:

# For HTTP transport
npx @modelcontextprotocol/inspector http://localhost:8000/mcp

# For stdio transport
npx @modelcontextprotocol/inspector python glances_mcp.py

Using curl (HTTP transport only)

# Test if Glances is responding
curl http://localhost:61208/api/4/status

# Test if MCP server is running
curl http://localhost:8000/mcp

Example Usage with Claude

Once configured, you can ask Claude questions like:

• "What's my current CPU usage?"
• "Show me the top 10 processes by memory usage"
• "What's the system load average?"
• "Check if my disk I/O is high"
• "Monitor network traffic"
• "Analyze my system performance and suggest optimizations"

Troubleshooting

Connection Issues

Problem: MCP server can't connect to Glances

Solution:

1. Verify Glances is running: curl http://localhost:61208/api/4/status
2. Check the GLANCES_URL in glances_mcp.py
3. Ensure no firewall is blocking port 61208

Empty Responses

Problem: Tools return empty or error responses

Solution:

1. Check Glances logs for errors
2. Verify the plugin exists: curl http://localhost:61208/api/4/pluginslist
3. Try accessing the specific endpoint directly via curl

HTTP Transport 404 Errors

Problem: Getting 404 when connecting to MCP server via HTTP

Solution:

1. Make sure to connect to /mcp endpoint: http://localhost:8000/mcp
2. Verify the server is running with ps aux | grep glances_mcp
3. Check server logs for startup errors

Claude Desktop Not Detecting Server

Problem: Server doesn't appear in Claude Desktop

Solution:

1. Check the configuration file syntax (must be valid JSON)
2. Use full absolute paths in the configuration
3. Restart Claude Desktop after configuration changes
4. Check Claude Desktop logs for errors

Development

Adding New Tools

To add a new tool, follow this pattern:

@mcp.tool()
def your_new_tool(param: str) -> dict[str, Any]:
    """
    Description of what your tool does

    Args:
        param: Description of the parameter
    """
    return make_request(f"/api/4/{param}")

Custom Glances Plugins

If you've added custom plugins to Glances, you can access them using:

get_plugin_data("your_plugin_name")

API Reference

For detailed information about the Glances API endpoints, see the Glances RESTful API documentation.

License

This MCP server follows the same license as Glances (LGPL-3.0).

Contributing

Contributions are welcome! Please:

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

Support

• Glances Issues: Glances GitHub Issues
• MCP Protocol: Model Context Protocol Documentation
• General Questions: Open an issue in this repository

Changelog

Latest Version

• ✅ Converted from httpx to requests library
• ✅ Added HTTP transport support via uvicorn
• ✅ Improved error handling for empty responses
• ✅ Added comprehensive documentation
• ✅ Fixed MCP endpoint configuration

Acknowledgments

• Glances by Nicolas Hennion
• Model Context Protocol by Anthropic
• FastMCP by Jeff Lowin