Continously increasing CPU Usage - how to find out why

I am observing a continously increasing CPU Usage of my HomeAssistant-System.

While it isn’t dramatic at the Moment, one could say as well that it increased by 100% within 6 Month :slight_smile:
And i really don’t know what causes that. So i want to find the culprit - but how can i do this?

I catch the CPU Usage (both % and load 5min) via the platform: systemmonitor . But i would need something more specific.

How can i tell which Addons, processes, Component increase in Usage?

Glances - Home Assistant could help you :wink:

I remembered that as well, but i think there is no Sensors for that Purpose:
Glances integration will add the following sensors if available in the platform:

For each detected disk (or mount point) the following sensors will be created:
    disk_use_percent: The used disk space in percent.
    disk_use: The used disk space.
    disk_free: The free disk space.
memory_use_percent: The used memory in percent.
memory_use: The used memory.
memory_free: The free memory.
swap_use_percent: The used swap space in percent.
swap_use: The used swap space.
swap_free: The free swap space.
processor_load: The load.
process_running: The number of running processes.
process_total: The total number of processes.
process_thread: The number of threads.
process_sleeping: The number of sleeping processes.
cpu_use_percent: The used CPU in percent.
sensor_temp: A temperature sensor for each device that provides temperature (depends on platform).
docker_active: The count of active Docker containers.
docker_cpu_use: The total CPU usage in percent of Docker containers.
docker_memory_use: The total memory used by Docker containers.
For each detected raid the following sensors will be created:
    raid_available: The number of available devices for the raid.
    raid_used: The number of devices used by the raid.

Or did i miss something terrible, which sensor(s) would you go for?

Well, it also gives you a live screen, where you can see the addons resources consumption…

1 Like

Well, it also gives you a live screen

I think that a Live screen might not be helpful if i want to observe increasing usage over a couple of month :wink: Anyway, i am not sure wether i could access that Screen with HAOS as System, too.

But yes, having that listing as sensors to keep track of would be perfect

So - no Possibility for CPU Sensors separated by Addons/Instances, in order to track their individual Usage over a longer Period? :frowning: