New releases quite hard on CPU ( on overview page )

Configuration Frontend
1

Hello all,
I’ve just finished my upgrade to 2025.12.5 hoping to fix the problem I’ve started to have after 2025.12.4 but with no success.
What’s happening is that, after both upgrade, the overview ( my overview I suspect ) page is hitting hard on my laptop CPU.
If I change to any other page I have in my installation, the CPU drops from 250 to almost 0, but when I came back to Overview then CPU jump to more than 200% ( which means is using more than 2 core of my PC ).
Has anyone never experienced similar behavior ?
What can I do ?

Thanks in advance.
Pierluigi

2

Not seeing anything like this. Which custom card are you using that is only shown on your overview page?

3

As custom card I use very few:
mushroom-chips-card
power-flow-card
mini-graph-card
mushroom-entity-card
mushroom-climate-card
mushroom-humidifier-card

Is not even such a complex page:

image
image1654×963 145 KB

4

It is probably not the card code that is causing it, since that would likely make an increase on the client instance instead.
If the increase is on the server, then it is more likely the data that is used in one or more of the cards.

5

Do you only use those cards on your overview page?

6

Yes. Those are the type of card I use on overview as I understand.
Do you want me to attach the whole yaml to get a look ?
I’m not very in in layout editing so I could be wrong.

7

So you don’t use mushroom cards on your other pages?

8

It doesn’t look I use them in other pages.

9

Try moving one section at a time to another page and see if you can narrow it down that way. Which section causes the other page to slow?

10

I will try this way.
It seems that no other page solws down.
I’ll try this later today ( time to work now :slight_smile: )
Thanks for the moment.
I’ll report back to you

11

I’ll first ask the obvious things: have you tried a full flush of the browser cache and/or opening the page in private mode? What browser do you use? How much memory does your laptop have?

What you could also try is start a new dashboard and try to recreate this as much as possible using standard tiles etc. who have grown in capabilities quite a lot over time.

I have heard before that people have had problems in the past with graphs causing redraws on each other in cascade. The cards you use seem quite ok but you never know until you try. If the cards have some kind of animation you can turn off, try to turn it off.

12

From my previous experience, it’s the moving dots in the energy card that generates this CPU usage. I switched to Power Flow card plus and deactivate moving dots and then CPU came back to normal.

2 Likes
13

If that is true I think it should be reported as a bug in Github.

14

I understood this might be a browser optimisation issue (related to GPU availability or not). So not related to HA.

15

You mean if you throw a GPU at it you won’t notice the bad implementation? It’s a freaking moving dot, not a bitcoin miner.

1 Like
16

I seem to remember the same thing happening with animated svg icons on a weather card.

17

If one card triggers a refresh on another card while animating then you can get caught in an avalanche if you have multiple cards triggering each other. That is why some people won’t notice and other people do. cards should not influence each other. People should also not need a GPU to hide the problem.

18

If it’s related to display, they might. My issue was on an old platform (Odroid C2) where Electron was not able to use the GPU acceleration, therefore the CPU had to manage itself the moving dot.

19

Which brings me back to: you should not need 2 cpu’s, both running 100% to move a dot.

20

It seems that removing the energy card fixed istantly the issue.
BTW the browser is a chrome ( not really new ), I’ve flushed cache, I have 16G of RAM.

Thanks all for ideas !
I will try to file a bug for this. Would it be fine on HA github or energy card has his own, AFAYK ?