just updated esphome to the 2026.02 update and also updated watchmon, upon restart the HA was actiing mega sluggish, so i pulled the power on my nuc and powered up again, i can ping the ip, i can access samba share through my pc, but the ui is just saying loading data and not doing much of anything, i did try to ssh into it but i cannot bloody remember my password for it, any way i can get in? ive tried to get in via my dns server and through the ip address direct but no joy
I haven’t updated ESPHome but I have updated Watchman, so you can rule that out as a cause. EDIT: apparently not. See posts below.
Additionally it is extremely unlikely that the ESPHome update is the cause too. It runs in its own container and only if you have updated all your ESP devices as well might they become unavailable if there was an issue with it (there are no major ones reported).
Never do this to a running server. This will likely cause more issues than it solves.
First try clearing your your web browser / app cache. Or try the private browsing mode of your web browser.
If that does not help then connect a keyboard and monitor to your server and use the CLI to view the logs.
i managed to get in, i deleted watchmon from the samba share, pulled power to the nuc, and then it loaded right up, my cpu temp was maxed out at 100c so it was hammering the cpu for some reason
@gazzaman2k Seems to be a bug with the new watchman update, I could only revert by removing Watchman altogether via SSH. Open Github issue is [here] (After update for the v0.8.1 2026-01-25 it briked my HA · Issue #222 · dummylabs/thewatchman · GitHub)
thx i knew it was watchmon as my temperture graph shows the moment i installed it my cpu went crazy hot
While not experiencing any issue, I can’t find Watchman in Devices & Services after the update. Which is odd. EDIT: I must have uninstalled the integration but not uninstalled it from HACS.
Spook pretty much does what Watchman does but raises repairs instead of generating a report.
I’m going to remove it.
Stop doing this. It is a sure fire way to corrupt your system.
1 Like
“I was driving down the autobahn doing 250kmh and saw somebody put on their brake lights ahead, so I moved the gear lever to Park position to slow down.
Now there is a pile of bolts underneath my car and it won’t go.
Stupid car. What to do?”
You’re doing the same thing, and especially during an update when CPU load is high and disk writes are intense.
DON’T!
A lot of system updates involve downloading and unpacking a lot of data in a controlled sequence, uninstalling an old version and replacing it with a new one, and following a set procedure with temporary intermediate steps to arrive at a known final solution. Expect maximum system load while doing that. Depending on your internet speed and system configuration, and how big the updates are, that can often take hours. Some people leave theirs overnight to do updates and check in the morning.
Watchman may have been getting confused during updates, hence the intense load. Haven’t looked at the algorithm behind the scenes, but it seems to have choked up some systems. The joys of open source software!
What to do? Start again with your update? Reinstall and restore from last known backup? Can you trust no update steps will be skipped as there are now orphan temporary files possibly hanging around? How to find these and clean them up?
Do you disable Watchman before updates, and then gingerly re-enable it again afterwards?
Nah it actually has serious issues. 🛡 Watchman - keeps track of missing entities and actions in your config files - #233 by pkscout
Sounds like something that should be co-developed with system updates, or something developed by one of the system developers if it works at such deep levels to fix underlying issues.
Like Spook?
Hmmm. So prior to all this info coming to light regarding the Watchman update issues, I had already hit update but not yet restarted HA. After reading about this I’ve now gone into the HA settings and deleted Watchman but again not yet restarted HA since I’m away from home. I wonder if I’ll escape the issue or if it’ll still smash my HA system.
What does Watchman do that shouldn’t be rolled into system functionality to improve it?
Ok to report problems by deep analysis, but how about focus on why they arise in the first place, and how to prevent that?
Prevention, rather than cure.
I’m guessing you will still have 100% system utilisation as the updates are not resource limited, the idea to get them over and done with as soon as possible, but Watchman won’t be there to meddle at this time.
I’d put money on everything being ok.
Hard agree on the unplugging the nuc. Don’t do that. It severely increases risk of corrupting something. Been there, done that. Surely, the nuc has a power button, yea? It’s still rather an uncouth way but holding that power button down for a hard shutdown is better than unplugging. It’s still an ungraceful shutdown, but it gives the system a chance to wipe buffers and park hardware, reducing risk of corrupting something. Don’t get me wrong, still a chance to mess something up, but unplugging is the worst.
Maybe restart in safe mode from SSH CLI and uninstall Watchman may be the safest and elegant workaround to try.