I ran ha dns options --fallback=false in terminal but with no luck. Are there other things I can attempt?
Roll back to OS 14.2.
v15.0 has been bad for a lot of people.
2 Likes
Thank you for the heads-up!
But I have to wonder; how would I know this prior to updating?
I always review the release notes and all the comments here before updating HA Core. They’re linked right from the Settings page in HA. But when I click the similar link in the HAOS update notification, I get a GitHub entry which is highly technical, but doesn’t really indicate what sort of issues I as a user should be looking for.
I noticed the same thing, which led me to this post. I must have updated to 2025.3 around 3/14.
I’m on 2025.3.3 currently, but I’m also still on HA OS 14.2, and haven’t updated to 15.0 yet…
Have you found any resolution(s) / workaround(s) to this?
Also, tangentially, can you provide any details / tutorials / etc… about the graph you posted? Looks like Grafana maybe? The only reason I started sniffing around was that I was seeing wonky behavior in HomeAssistant all over the place, and saw that HA’s CPUs were maxed out in Proxmox. Having a graph would have made some of my sleuthing much quicker. 
Just another data point: When I updated from Core 2025.2.5 to 2025.3.4 I saw absolutely no change in CPU utilization.
I’m still waiting to decide whether to install 15.0. This thread reports that the problems may be specific to certain hardware configurations, and I believe them
But that thread doesn’t mention mention CPU utilization at all. The good news is it reports that about one-third of all installations reporting statistics have updated. This gives me some confidence, but I’m quite sure my hardware is in the minority.
Yep, it’s InfuxDB + Grafana. It looks like both are available as HAOS Addons, but I have mine running on UnRAID and just point telegraf on proxmox to it, as well as the HA integration for InfluxDB for some individual sensors that I want long-term stats for.
It’s great being able to go back and look at things over several months. I’ve had it set up for a long time and I know that it’s about time to try to upgrade the stack because I’m still using an older Grafana and InfluxDB v1… but it’s one of those things where it isn’t causing a problem so I keep pushing it off.
1 Like
I don’t think it’s InfluxDB and Graphana, as I have the increased CPU usage, but none of those two addons.
I do run HAOS in a VM in proxmox.
My reply about Influx + Grafana was responding to the user that asked about how the screenshot of the graph that I posted was created… I wasn’t saying that the high CPU usage was caused by Influx + Grafana, just that it’s what I used to produce the graph showing the Home Assistant docker container’s CPU usage.
I found the problem on my end. It was CompreFace.
I disabled it and now load is back to normal.
ok. Problem is solved with a newer update too.
1 Like
15.2 fixed it for you? I stumbled upon the same issue the other day and disabled Compreface for the time being. Just updated to 15.2 and started Compreface back up and still have the same error loop with CPU skyrocketing 
Tried starting CompreFace again on 15.2. Problem persists.
possibly independent problems then, as I don’t use those addons.