High CPU Utilisation - Help / Advice?

Hi There, I’ve 2 instances of Hassio - one at home and another at a holiday house. The hardware setup is pretty much the same although one at home only has a 32G sdd (+zwave stick) and the holiday one 64G sdd has no directly attached devices. I’ve got the same addon’s loaded across both instances (PiHole/NodeRed/Nginx Proxy, etc) although some of the devices and brands are different.

There are no obvious hints in the HA logs, however the “top” command for the home instance shows;

Whilst the top command on the holiday instance shows;

Any advice on trying to work out what is going on? The holiday instance is running with a consistently high CPU temperature >85 Celcius so I’d like to address before the hardware fails…

Disable addons one by one looking for the culprit.

Or

Install the community Portainer addon and use it’s stat’s function to check out which container is hogging the CPU. You will have to do this with protected mode ‘off’ for the addon.

Thanks for the quick response @tom_l. I’ve got Portainer running on both instances but please forgive me for asking stupid questions on account that I am a hobbyist and not a techy, where do I look for the stats? I’ve tried my best friend google but drawn a blank.

FYI - the dashboard throws up this;

First off, you have to do this:

Tun of protection mode in the addon page. Then go to the Portainer tab and do this:

By default all Hass.io managed containers are hidden from Portainer.
This is recommended since fooling around with Hass.io managed containers
can easily lead to a broken system.

  1. Go into Portainer → Settings → Hidden containers:
    Delete the listed hidden labels (io.hass.type labels).

You should then see the number of containers (in your image above) be above zero.

You may have to restart the addon, I can’t remember.

Anyway assuming you can now see the number of containers click on that containers icon and you should see a list of your containers.

Click on settings (top right of that page) and make sure stats is ticked.

Now there should be a graph symbol next to each container in the list (under the quick actions column). Clicking on a graph symbol will bring up the stats page for that container. Wait for it to populate (no history is stored).

1 Like

Thank you @tom_l - you are a superstar! Portainer makes a lot more sense to me now.

A quick skim of the stats identified motioneye as the culprit and stopping the container has reduced the CPU load and the temperature has dropped to :smiley:

I now just need to work out now why motioneye is causing a problem!

1 Like