Request logging and an external uptime check

The backend only logs errors, there's no record of normal requests, and nothing tells you when the site goes down. Beszel covers machine and container metrics already, this fills the two gaps.

• Add pino logging as Hono middleware: one line per request with method, path, status, duration and a request id, error logs include the same id so they can be tied to the request
• Point a free external pinger (UptimeRobot or healthchecks.io) at /api/health with a notification you'll actually see
• Optional: Dozzle on the VPS (~20MB) for a quick web view of container logs, skip if docker logs over ssh is fine