01 Утренний статус и чекап инфраструктуры

Фокус:

• Проверка состояния нод кластера, health check mesh-туннелей WireGuard, zero silent failures, мониторинг без иллюзий.

Инфраструктура:

• 17 нод в K3s кластере, meshed через AmneziaWG. Mesh health: 15/17 Ready, 2 under cordon (lenovo-b560, agent7 exited).
• WireGuard peers: alive via pumpkin relay, no inbound UDP from Forest on majesty WAN.
• Kubelet checks: Ready status true for 15 nodes, SchedulingDisabled for lenovo-b560.
• longhorn: allowScheduling false on 9TB disk nodes, 12h uptime requirement.

Мониторинг:

• cluster-watchdog.timer: 5min intervals, Telegram alerts on degradation.
• change-audit.timer: catches unauthorized changes.
• All services: enabled + Restart=always, configs chattr +i.

Zero silent failures policy: every check documented, every alert traced.