How do you tell everyone at home when an automation is down?

When a hub or integration goes offline, the person who built the setup may notice first, while everyone else just sees a light or thermostat behaving oddly. I’m interested in the handoff: what deserves an immediate alert, what can wait for a shared status view, and what manual fallback should be easy to find? For a noncritical lighting rule, I’d favor a quiet status with a clear wall-switch fallback; for access or heating, I’d want a more visible notice. How do you keep those alerts useful without training the household to ignore them?

4 Likes

I think Basic Home Guard can tell you if a device is offline. Other than that, I don’t think there’s a way in the Aqara app to tell an automation is offline (at least automatically).

In my home we rely on the user’s reports of any issues. So basically - my wife tells me the light is broken :joy:

We use wall switches and buttons as a fallback, but they are just convenient (more convenient than using the phone to open the app, find a button etc etc, especially when one of your hands is full or you don’t have pockets in your current clothes). So they are also used on daily basis just to turn on/off the light manually (or to change the lighting setting)

6 Likes

That distinction helps: a device-offline alert and a rule that silently stops working are different problems. Keeping the wall controls useful every day also makes the fallback familiar when a rule fails. Is a household report your main signal for silent failures, or have you found a low-noise way to catch them earlier?

1 Like

This is the ultimate test of the “Spousal Acceptance Factor”. If you spam the family with non-critical alerts, they will just mute everything and you will still get the dreaded “the house is broken again” text.
Here is how I structure the handoff to keep the peace:
Direct the noise to the admin: Route push notifications for offline hubs or integration drops only to your own phone. The rest of the household does not need to know a Thread border router dropped off the network, they just want the lights on.
Subtle visual cues: For non-critical offline states, use the LED ring on an Aqara M3 hub or flash a specific accent light. A quiet red pulse lets you know the mesh is acting up without bothering anyone else.
Hardware fallbacks are non-negotiable: Rely on Zigbee binding or physical relays. If your coordinator or Home Assistant instance goes down, an Aqara wall switch must still operate the bulb physically. Never make an app the only way to turn on a room.
Save the sirens for disaster: Water leaks, freezing temps, or security breaches get pushed to everyone’s devices.
Basically, if the house is flooding, wake everyone up. If a smart plug falls off the mesh, just let the resident nerd quietly fix it before anyone notices

4 Likes

For the silent failures James mentioned, I’d monitor the expected result, not just whether the hub is online. In my Home Assistant setup, presence drives lighting and the AC is controlled through an API. For that kind of rule, I would start a health check only when the normal conditions are met, allow a settling period, then compare the requested target with whatever actual feedback the device provides. An intentional manual override should not count as a failure. Missing or unavailable presence data should be a separate fault, not silently treated as an empty room.

Home Assistant’s traces help distinguish a failed condition from an action error when an automation ran. They still do not prove that a physical device responded, especially if the integration only reports an assumed state after sending a command. I would use that distinction in the status message, for example ‘Living room lighting could not be confirmed’ instead of ‘hub offline’. It makes the fallback more useful because the household knows which function is affected, while the detailed trace stays with the person fixing it.

2 Likes

The requested-versus-confirmed distinction belongs in the household status, not just the admin trace. If an integration echoes the command without device feedback, I’d label the result ‘unconfirmed’ and show when the last real device state was seen; a delayed check could then separate a transient stale state from a persistent failure. That should also keep a deliberate manual override out of the fault queue.

1 Like

Hey there.there is no app that could tell you if an automation is offline.the only way to know without using an app is for example set an notification when that specific device was turned ON or off by that automation.but to make sure that doesn’t happen you need an strong thread or zigbee connectivity and make sure to make automation locally.i never bother to think that an automations goes offline because all my automations are made via M3 and are local, i have an strong connectivity on my WiFi also in thread and zigbee and if my WiFi goes off then the automations are still tuning without an issue.

1 Like