The idea is this: schedule 'checkups' regularly, either on some schedule (like cron) or after events are triggered. Those checkups can try to assert a particular state, and if they aren't right, log it (ie report it as an error to something like rollbar), and optionally, if there's a known fix, try to fix it.
Some ideas I've had so far:
every 15 minutes, check the current position of the sun (sun.sun), and check exterior lights to be on/off as expected. they can be fixed
every so often, check for Hue service to be present (it's not been coming up at startup for some reason), and check that all bulbs are online
every so often, iterate over groups, and check that all the entities in them actually exist
I've been having trouble with some of my automations not firing, particularly around sun triggers. I saw a talk at RailsConf 2018 about The Doctor Is In: Using checkups to find bugs in production
The idea is this: schedule 'checkups' regularly, either on some schedule (like cron) or after events are triggered. Those checkups can try to assert a particular state, and if they aren't right, log it (ie report it as an error to something like rollbar), and optionally, if there's a known fix, try to fix it.
Some ideas I've had so far:
sun.sun), and check exterior lights to be on/off as expected. they can be fixed