After a brief outage, hubs, routers, cameras, and cloud services do not all come back at the same time. I avoid letting noncritical automations immediately resume at full speed, because a sensor event during partial recovery can create missed clips or duplicate notifications.
For one test outage, I record:
- when router internet access returns;
- when each Aqara hub shows online;
- when one Zigbee sensor and one Thread or Matter device respond;
- when a camera live view and recording action work;
- whether queued automations fire late.
Then I add a short recovery condition: only re-enable the dependent automation after the hub and target device have stayed available for five minutes. I also keep one manual scene that runs the camera or notification path once before normal schedules resume. This turns “everything looks online” into an end-to-end check instead of trusting the first green status icon.
3 Likes
@jamesfengyuyang This is a really thoughtful approach to managing post-outage recovery! Your staged verification—tracking router, hub, Zigbee/Thread devices, and camera functions separately before re-enabling automations—addresses a pain point many users overlook. The five-minute stability window and manual test scene are smart safeguards against the “green icon but broken pipe” scenario.
A couple of things worth noting from the ecosystem side: Aqara hubs (M3, M2, G3, etc.) do support proxy hub functionality where certain Zigbee devices like the Wall Outlet H2 EU or Dual Relay Module T2 can maintain basic local automations even if the main hub is down. However, this only covers simple on/off actions, so your caution about complex automations firing during partial recovery still applies.
Your logging method also highlights something useful for the community—having timestamped visibility into which layer recovers when. If you ever want to share a template for that checklist or compare notes on how different hub models behave during staggered recovery, that’d be valuable content.
Have you noticed any patterns in which devices tend to lag behind the hub’s “online” status? Thread border routers and camera hubs sometimes have additional boot sequences that aren’t immediately obvious from the app indicators.