yaml automations to send notifications when the sensor enters/exits running. Its conditional on an input_boolean being on, so you can test without being distracted with notifications.
automation:
- alias: "appdaemon started"
trigger:
platform: state
entity_id: sensor.appdaemon_status
to: 'running'
condition:
condition: state
entity_id: input_boolean.notifications
state: 'on'
action:
service: notify.notify
data:
message: Appdaemon has started
title: HomeAssistant Error
- alias: "appdaemon stopped"
trigger:
platform: state
entity_id: sensor.appdaemon_status
from: 'running'
condition:
condition: state
entity_id: input_boolean.notifications
state: 'on'
action:
service: notify.notify
data:
message: Appdaemon has stopped
title: HomeAssistant Error
the idea is nice, but i dont think im gonna use it.
the only moment that appdaemon stopped working in all the time i used it is when my RPI stopped working.
at that moment HA is also stopped so thereās no notification then
but ill keep it in mind. maybe in the (near) future i will splitt up the machine that runs HA and Appdaemon again, and then i can let appdaemon send a notify when HA is down and HA send a notify when appdaemon is down.
I agree with @ReneTode in that the execution of this might not be what Iām looking for either, but the idea behind it is fantastic! Working with outside sources of information, that is, not specific to the Home Assistant service and entities within Home Assistant itself, is a GREAT idea.
That said ā how often do you find AppDaemon stopping? If the service is setup correctly, this should not happen.
For what itās worth, you can run AppDaemon on multiple computers, all simultaneously. You could even have each instance āpingā the other to make sure itās still responding. If itās not responding, have the fallback ātake overā in case of listeners.
I would also say I have found it to be very reliable but I am biased, Iām gratified that others spoke up on the subject:
I still think itās an intriguing idea though - once or twice I have stopped AppDaemon on during tests and wondered why everything stopped working ā¦
in the future i would consider running HA on 1 device and AD on another.
but both installed on both devices.
with the crosscheck you then always can use 1 or the other to start the backup version on the same machine.