Observium MCP Server

Observium MCP Server

A Model Context Protocol (MCP) server that exposes Observium CE network monitoring data to LLMs like Claude, enabling natural language queries about device status, network traffic, sensor readings, alerts, and historical trends.

Features

• Device Management: List and query monitored devices with status, uptime, and hardware info
• Network Ports: View interface status, traffic rates, and utilization
• Sensors: Access temperature, voltage, frequency, and other sensor data
• Alerts: Query active and historical alerts with summaries
• Trends: Retrieve historical metrics from RRD data (load, CPU, memory)

Requirements

• Python 3.10+ (required by MCP SDK)
• Observium CE installation with MySQL/MariaDB database access
• Access to Observium's RRD data directory (for trend data)
• rrdtool and librrd-dev packages (for RRD data access)

Installation

1. Clone the repository

git clone https://github.com/kdesch5000/observium-mcp.git
cd observium-mcp

2. Create a virtual environment

python3 -m venv venv
source venv/bin/activate

3. Install dependencies

pip install -r requirements.txt

4. Configure environment

Copy the example configuration and edit with your Observium database credentials:

cp config.example.env .env

Edit .env with your settings:

# Database connection (use SSH tunnel if remote)
OBSERVIUM_DB_HOST=localhost
OBSERVIUM_DB_PORT=3306
OBSERVIUM_DB_NAME=observium
OBSERVIUM_DB_USER=observium
OBSERVIUM_DB_PASS=your_database_password

# RRD data path (on the Observium server)
OBSERVIUM_RRD_PATH=/opt/observium/rrd

# Optional: SSH configuration for remote RRD access
# If MCP server runs on a different machine than Observium,
# set these to enable SSH-based RRD file access for trend data
# OBSERVIUM_RRD_SSH_HOST=observium.example.com
# OBSERVIUM_RRD_SSH_USER=pi
# OBSERVIUM_RRD_SSH_PORT=22

Usage

With Claude Code

Add to your Claude Code MCP configuration (~/.claude/claude_code_config.json):

{
  "mcpServers": {
    "observium": {
      "command": "python",
      "args": ["-m", "observium_mcp.server"],
      "cwd": "/path/to/observium-mcp/src",
      "env": {
        "OBSERVIUM_DB_HOST": "localhost",
        "OBSERVIUM_DB_NAME": "observium",
        "OBSERVIUM_DB_USER": "observium",
        "OBSERVIUM_DB_PASS": "your_password",
        "OBSERVIUM_RRD_PATH": "/opt/observium/rrd"
      }
    }
  }
}

With Claude Desktop

Add to your Claude Desktop configuration:

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

{
  "mcpServers": {
    "observium": {
      "command": "python",
      "args": ["-m", "observium_mcp.server"],
      "cwd": "/path/to/observium-mcp/src",
      "env": {
        "OBSERVIUM_DB_HOST": "your_observium_host",
        "OBSERVIUM_DB_NAME": "observium",
        "OBSERVIUM_DB_USER": "observium",
        "OBSERVIUM_DB_PASS": "your_password",
        "OBSERVIUM_RRD_PATH": "/opt/observium/rrd"
      }
    }
  }
}

Standalone

cd src
python -m observium_mcp.server

Available Tools

| Tool | Description | |------|-------------| | list_devices | List all monitored devices with status | | get_device | Get detailed info for a specific device | | list_ports | List network interfaces for a device | | get_port_traffic | Get traffic stats for a specific port | | list_sensors | List sensor readings (temp, voltage, etc.) | | list_alerts | List active or historical alerts | | get_alert_summary | Get alert count summary | | get_trends | Get historical metric data | | list_available_metrics | List available RRD metrics for a device |

Example Queries

Once connected, you can ask Claude questions like:

• "What devices are currently down?"
• "Show me the temperature sensors on the main switch"
• "What's the CPU load trend for the firewall over the past week?"
• "Are there any active alerts?"
• "Which ports on the core switch have errors?"
• "What's the uptime of all my Linux servers?"

Remote Access

If your Observium instance is on a remote server, you have several options:

Option 1: SSH Tunnel

# Create SSH tunnel to forward MySQL port
ssh -L 3306:localhost:3306 user@observium-server

# Then configure .env with localhost
OBSERVIUM_DB_HOST=localhost

Option 2: Install on Observium Server

Install the MCP server directly on the Observium host and configure Claude to connect via SSH.

Option 3: Network Access

If MySQL is accessible on the network (not recommended for security):

OBSERVIUM_DB_HOST=observium.example.com

Security Considerations

• Database credentials are stored in .env which is gitignored
• The server only performs SELECT queries (read-only)
• Consider using a read-only MySQL user for additional safety
• RRD access is read-only via rrdtool fetch

Creating a Read-Only Database User

For additional security, create a dedicated read-only MySQL user:

CREATE USER 'observium_mcp'@'localhost' IDENTIFIED BY 'secure_password';
GRANT SELECT ON observium.* TO 'observium_mcp'@'localhost';
FLUSH PRIVILEGES;

Troubleshooting

Connection refused

• Verify MySQL is running and accessible
• Check database credentials in .env
• Ensure the MySQL user has SELECT permissions

No RRD data / Trend tools failing

• Verify OBSERVIUM_RRD_PATH points to the correct directory on the Observium server
• If running MCP on a different machine, configure SSH access:
• Set OBSERVIUM_RRD_SSH_HOST to the Observium server hostname
• Set OBSERVIUM_RRD_SSH_USER to a user with SSH access
• Ensure SSH key-based authentication is configured
• Ensure rrdtool is installed on the Observium server (for remote) or locally
• Check file permissions on the RRD directory

Module not found

Make sure you're running from the src directory or have installed the package:

cd /path/to/observium-mcp/src
python -m observium_mcp.server

Contributing

Contributions are welcome! Please feel free to submit issues and pull requests.

License

MIT License - see LICENSE for details.

Acknowledgments

• Observium - Network monitoring platform
• Model Context Protocol - The MCP specification
• Anthropic - Claude and MCP development