Health Monitor FAQs

Question 1

Where is my server health data stored?

Accepted Answer

All health history, including latency trends and uptime data, is stored locally in a SQLite database on your machine. No external monitoring service is required, ensuring data residency and privacy.

Question 2

What makes Health Monitor unique compared to generic uptime tools?

Accepted Answer

Health Monitor speaks the Model Context Protocol (MCP) directly, calling `list_tools` to ensure your servers are genuinely ready to serve requests, not just responding to HTTP pings. It also detects tool schema version drift, which generic monitors cannot do.

Question 3

How does Health Monitor discover my MCP servers?

Accepted Answer

It automatically discovers MCP servers by reading existing client configuration files from tools like Claude Desktop, Cursor, and VS Code. You can also manually register additional servers.

Question 4

Can I view historical performance and latency trends for my MCP servers?

Accepted Answer

Yes, Health Monitor stores historical latency data and computes p50/p95 percentiles on-demand, allowing you to identify regressions before they impact users. It also generates an HTML dashboard with uptime sparklines.

Question 5

Does Health Monitor detect when my server's tool schema changes?

Accepted Answer

Yes, a key feature is its ability to detect 'version drift'. It hashes tool schemas on each probe, comparing them against previous versions to alert you when a server update changes its exposed tools.

Health Monitor

Health Monitor

主な機能

ユースケース

主な機能

ユースケース