Periodic heartbeat log + nudge toward running as a service

[fayerman-source]

Observed: running via a foreground npm start. Between motion events there is no signal the process is still alive, and closing the terminal (or a WSL restart) stops recording silently.

Ideas:

• Emit a periodic heartbeat/status line (e.g. hourly): uptime, cameras watched, clips recorded today, time of last event. Makes long-running health visible at a glance.
• Strengthen the README to recommend the systemd/pm2 units in deploy/ for 24/7 use (survives logout/reboot) over a foreground terminal.

Observability/operability, not a bug.