Hi all,
The last week I have had several occurrences of HA stop working. That is, HA doe work, but none of the services can be called anymore. I see that the services are called, but, for example, my switches remain turned of.
But it applies to all the technologies, incoming and outgoing. MQTT presence detection stops receiving updates, ifttt integration stops working, etc.
I have not found any error in the logs. Does anyone else experience the same problem? I’m running HA on a rpi3.
Ok, wel… never mind.
After some tinkering I found out that the API layer becomes unresponsive. I don’t know the reason yet, but right now I have a script running every minute (through cron) that calls /api/states to verify if HA is up. If there is a timeout, it kills the process and restarts HA.
Strangely the default call to /api/ to check if it is up always returns the ok message, even though the rest of the API is unresponsive.
Apparently the problem was outside of Home Assistant. It seems to be a reconfigured Fail2Ban installation. After removing that the amount of restarts due unresponsiveness were drastically reduced. Not quite sure on the how or why part, but it is better than nothing.