Health Monitor

Monitors MCP server health by probing the protocol directly via list_tools calls, with auto-discovery, latency tracking, and version drift detection.

This tool monitors MCP server health using native protocol probing instead of simple HTTP pings. It auto-discovers servers from Claude Desktop, Cursor, and VS Code configuration files, then periodically calls list_tools on each to verify genuine protocol readiness. Health history is stored locally in SQLite with p50/p95 latency percentiles, and a self-contained HTML dashboard provides uptime sparklines per server.