It is useful to have ‘health’ sensors for integrations to be able to observe if they are operating normally.
Now there is the risk that automations take wrong conclusions when entries are unavailable. It can make a big difference if just that particular device is down, or the whole integration is down. Especially relevant for hub type of integrations (ZHA, MQTT, Hue, Z2M, …).
As a work around, I created template sensors to reflect integration issues if all entities owned are lost. A little brute force, and integrations themself should know best.
I don’t understand… so what do you actually propose?
I propose that integrations publish a health/status sensor. These sensors should refect the status of the integration. e.g:
- initializing ( at startup or reload )
- operational (e.g. Connected with hub and hub is ok)
- error (e.g ZHA dongle not connected)
- disabled
1 Like