FleetWatch

Real-time uptime monitoring for MCP servers. Register a server, FleetWatch pings it on a schedule, and tells you the moment it goes down, gets slow, or comes back.

None of the other tools in this portfolio do this: mcp-schema-watch catches breaking schema changes, contextpulse watches context/token health, skillcheck lints SKILL.md files, Swarm Trace Viewer debugs finished runs, PR Blast Radius flags risky pull requests. FleetWatch answers the one question none of them do: is the server actually up, right now.

How it works

• Register an MCP server's URL.
• A BullMQ repeatable job pings it every HEALTH_CHECK_INTERVAL_MS (default 60s) with a tools/list JSON-RPC call.
• Response time and success/failure are recorded. Status becomes up, degraded (slow), or down.
• On any status change, an alert is written and pushed live over WebSocket. NotificationService currently logs alerts — wire in Slack/email there when ready.
• The dashboard shows live status per server with no page refresh.

Verified end to end: register → immediate check → status flips to down on an unreachable URL → alert recorded with the failure reason.

Run it locally (Mac, using your existing Homebrew Postgres/Redis)

brew services start postgresql
brew services start redis

psql postgres -c "CREATE USER fleetwatch WITH PASSWORD 'fleetwatch';"
psql postgres -c "CREATE DATABASE fleetwatch OWNER fleetwatch;"

cd api
cp .env.example .env
npm install
npm run start:dev             # http://localhost:3010

cd ../web
cp .env.local.example .env.local
npm install
npm run dev                   # http://localhost:3000

Prefer containers instead? docker-compose up -d from the repo root starts Postgres + Redis with the same credentials, skip the brew services / psql lines above.

What's stubbed for v1

• Notifications log to console instead of Slack/email — api/src/alerts/notification.service.ts.
• No auth yet — fine for local/personal use, add before deploying anywhere public.
• synchronize: true on TypeORM creates tables automatically for local dev. Swap for real migrations before production.

Stack

NestJS, TypeORM, BullMQ, Postgres, Redis, Socket.IO — Next.js, Tailwind.