System Monitor MCP Server for Linux

A Model Context Protocol (MCP) server implementation in JavaScript, designed for AI agents to monitor and interact with Linux systems. This server provides comprehensive metrics including CPU, memory, disk, network, and process data through both Stdio and secured HTTP/SSE transports.

Features

• Dual Mode: Runs as a standard MCP server (Stdio) or a secured HTTP server (--http).
• Performance Optimized: Uses TTL caching for static hardware info and parallel sensor polling.
• System Information: Hostname, OS details, kernel version, uptime.
• CPU Monitoring: Usage %, frequency, cores, brand, and temperature.
• Memory Monitoring: RAM and swap usage, real-time availability.
• Disk Monitoring: Storage usage, file system types, and mount point filtering.
• Network Monitoring: Interface statistics, traffic data, and error tracking.
• Persistent Metrics: Background persistence to history.json with a rolling 24-hour window.
• Proactive Alerting: Real-time threshold monitoring (CPU, Memory, Disk, Temp) with immediate console warnings.
• Advanced Process Tracking: Filter, sort, and limit process lists by resource consumption.

Security

HTTP Authentication

When running in HTTP mode (--http), you can secure the endpoint using an API key.

1. Set the MCP_API_KEY environment variable:

   export MCP_API_KEY="your-secure-token-here"
   node src/index.js --http

2. Clients must then provide the key in the x-api-key header or as a query parameter:

   • Header: x-api-key: your-secure-token-here
   • Query: http://localhost:57996/sse?apiKey=your-secure-token-here

Note: If MCP_API_KEY is not set, the HTTP server will default to public access (local network only recommended).

Usage

MCP Mode (Default / Stdio)

Used by MCP-compatible clients like Claude Desktop or OpenCode.

node src/index.js

OpenCode Configuration (~/.config/opencode/opencode.json):

"system-monitor": {
  "type": "local",
  "command": [
    "node",
    "/var/home/alexa/.gemini/extensions/system-monitor-js/src/index.js"
  ],
  "enabled": true
}

HTTP Server Mode

node src/index.js --http

The server listens on port 57996 by default.

Tools

• get_system_info: General OS/Kernel details.
• get_cpu_info: Load, specs, and temperature.
• get_memory_info: RAM/Swap metrics.
• get_disk_info: Partition usage. Parameters: mountPoint (optional).
• get_network_info: Traffic and interface stats.
• get_processes: Active processes. Parameters: limit (number), name (filter), sortBy (cpu_usage, memory_usage, priority).
• get_system_metrics: Consolidated system health snapshot.
• get_metrics_history: Retrieve historical performance trends. Parameters: limit (number of minutes).
• get_battery_info: Power status and battery health.
• get_usb_devices: Connected hardware list.

Validation Status

Linting Status

Status: ✅ Passing
Last Updated: 2026-01-25 12:03:39 UTC
Summary: 0 errors, 0 warnings

• 2026-01-25
• HTTP Authentication: Added support for MCP_API_KEY environment variable to secure the SSE/HTTP endpoint.
• Input Validation: Implemented Zod schemas for all tool parameters to ensure strict input handling.
• Code Consolidation: Merged mcp_server.js into index.js for a cleaner, unified codebase using the official MCP SDK for both transport modes.
• Static Caching: Added a TTL-based cache (1 hour) for static system data (OS info, CPU hardware) to reduce redundant system calls.
• Parallel Data Fetching: Optimized all data-gathering functions to use Promise.all for parallel hardware sensor queries.
• Renamed Identity: Rebranded the server from mcp-system-monitor-js to a simpler system-monitor.
• Persistent Metrics: Implemented background persistence to history.json with a rolling 24-hour window.
• Proactive Alerting: Added real-time threshold monitoring (CPU, Memory, Disk, Temp) with console warnings.
• Trend Analysis: Introduced get_metrics_history tool to retrieve performance history.
• Enhanced get_processes: Added support for limit, name filtering, and sortBy (cpu_usage, memory_usage, priority).
• Enhanced get_disk_info: Added mountPoint filtering support.
• Aggregated Metrics: Improved get_system_metrics tool to return a consolidated, prioritized snapshot of system health.
• Auto-Update: Added non-blocking auto-update script that checks for newer versions on startup.