- nginx stopped on VM1 (fault injected)
- health probes fail -> confirming
- still down after 6 s window -> fail over
- ssh pi4 'systemctl start nginx'
- backup is serving the site [ok]
Automatic failover for a small network. Health probes catch a dead server, SSH starts the standby and DNS moves the traffic, with nobody touching a keyboard.
- Failover
- 14.0 s on real hardware, target under 30 s
- Availability
- 99.97% in testing, target 99.9%
- Detection
- Probes every 5 s, then a 6 s confirmation window
- Hardware
- Raspberry Pi 5 controller, Pi 4 hot standby
- Team
- 5 students, with industrial partner Ghaymah Systems



