I’ve been using, happily, Home Assistant Green for about two years now without really any issue.
However, twice in the past three months, my instance has gone unreachable, with the yellow LED on the device blinking in a steady pattern. The docs only mention that a heartbeat pattern means it’s operational. I’m curious what this steady pattern on the yellow LED could be, maybe that will help me diagnose the issue, because of course it has happened both times at night as I’m trying to turn off all my lights.
I could probably look through logs to see what happened before the system became unreachable. If you think that’d be a helpful venture, I could do that, too.
When it happens again then connect a monitor and a keyboard to it and see if there is an output and if it reacts to kets being pressed.
Also check any log output on the screen as this can contain important information in diagnosing the issue.
I run a loaded Green too, so I’ve chased this once for myself before. In my case it was the power source, and replacing it solved the issue. I also built a little log regimen along the way that might help others, although I wouldn’t call it “easy to use for everyone”.
Power: My Green did go away at random, always at night, goin fully unreachable with no output. But I couldn’t plug it to a monitor/screen because of the location of the device. An under-spec or aging 12v supply will hold all day and then drop the board once the evening load piles on. Try the official supply, or a known-good 12v with a decent cable. Swap it even if you’re already on the stock one, since they age, and for me that replacement was the fix.
Logs: If power turns out fine, the logs are the next problem. HAOS tends to lose the pre-crash trail on reboot. Reaching the previous boot’s host logs has been broken upstream at times (operating-system #3817), and the plain home-assistant.log was dropped for OS installs (#1527) only to come back again as opt-in later on. All that made me run a small script on the box that copies the previous boot’s logs to persistent storage so a restart can’t wipe them: Github Gist. Acknowledging however that the better long‑term fix for the community probably belongs upstream in the project.
One honest caveat: a real power hang usually leaves nothing in the logs anyway, and that gap is itself the sign it dropped below Home Assistant rather than crashing inside it. I’d start with the power supply and only lean into additional logging script installations if you are comfortable with that kind of tinkering.
The option was reinstated a month (or 2) later. Raw logs from before the restart within the UI were preserved, so you can simply keep scrolling up the logs once access to HA was restored.
No idea why you would continue to run your own script when you only need 3 commands to restore the functionality Common tasks - Home Assistant Operating System - Home Assistant
I’m using the official power supply (always have). Any chance it could be related to the surge protector that is running on? (Everything else on that circuit is fine though (router & Eufy hub)
I’m assuming based on these comments that the HA Core logs are the ones I need to check out for this sort of problem?
Also (sorry, can’t figure out on mobile how to reply with a quote) I’m assuming based on these comments that the HA Core logs are the ones I need to check out for this sort of problem?
For me, replacing the power supply (PSU) stopped the outages.
The snapshot was part of that debugging journey to the PSU, motivated by HA’s documented gaps and quirks about host level logging, and changes in Core application logging.
