Monitoring

MCPHub monitors your MCP servers through periodic health probes and real-time WebSocket updates.

How it works

1. Health probes — MCPHub sends a probe request (MCP initialize call) to each registered server
2. The response time, HTTP status, and any errors are recorded
3. Server status updates to healthy, degraded, or down
4. The dashboard reflects current status in real time via WebSocket

Triggering probes

MCPHub is deployed on Vercel's Hobby tier, which supports one background cron per day. Rather than waiting for the cron, you can trigger probes on demand from the dashboard:

• Single server — open the server's detail page → Probe Now
• All servers — click Run All Probes on the Servers page

→ Health probes

Viewing history

Each server has a timeline of probe results going back 30 days, plus an uptime calendar view.

→ Health timeline

Live updates

The dashboard updates in real time via WebSocket — status dots change, alert toasts appear, and stats cards refresh without a page reload.

→ Real-time dashboard