Hello!
Can anyone help me find out why my home assistant supervised instance running on Raspberry P4 8Mb inside an Argon ONE M.2 Case, installed on Debian 11 is running high on CPU Temp at 2:25 a.m. ? Any idea what process is kicking in at the same time everyday?
Where should I look to find the answer?
The database usually purges at 04:12 local time. Is your browser 30 minutes behind local time?
I do not think my browser would be 30 minutes behind local time…
I looked closely and the process is starting at 2:00 a.m. and temperature peaks at 2:25 a.m.
Maybe it would be the database purge, how can I check this out?
It’s not the database purge. That is hard coded for 04:12 local time.
https://www.home-assistant.io/integrations/recorder/#auto_purge
Any other ideas as to what might use so much cpu power and raise temperature at 2:00 a.m. ??
Do you have an auto backup scheduled daily?
Look in your logbook to see if any automations are running at that time.
I guess this is where you are suggesting me to look at:
But I did reset my host at 10:12 a.m., I can’t look before that. I will keep an eye later tonight at 2:30 a.m. and update you then?
Or was you referring to another location in HA for logs?
No that is the log. I mean logbook. Click this:
Thank you Tom,
I would say not much is happening around 2:00 a.m. that could trigger so much cpu ressources:
I am still confused as to what may cause sudden temperature rise everyday around 2:00 a.m.
Are you sure your timzone and time are set correctly?
And you have not set up any automatic backup automation or system?
To test if it is the database purge you could add this to your configuration.yaml file:
recorder:
auto_purge: false
Then restart home assitant.
This will turn off the database purge for one night, just to see what happens.
Make sure to remove it and restart the next day.
Timezone is correct, I am in Quebec, same timezone as Toronto:
I will add recorder:
auto_purge: false
and see how HA acts this night
I will update you later about it.
The same happened this early morning, same time as usual:
And while looking at the log, I can’t find anything that could trigger so much ressources on the cpu in the interval between 2:00 a.m. and all the way up to max cpu temperature.
Could it be something running under the hood in Debian 11 that would run every night at 2:00 a.m.?
I have respected all requirements to run HA Supervised, but could it be that Debian is running a process I am not aware of? If so, where would the log be? How can I consult it?
What is the cpu and memory doing at that point, are there spikes there as well? Could it be something environmental, like physically warming it?
Glances(there is an addon) is able to show processes active and the cpu/mem use. I don’t know if you can record it, meaning you may need to be up watching when it happens. Or set up a camera to record the screen.
Yes, I’d recommend to check the crontabs of your system. Looks like there’s a job outside of HA running at 2 o’clock…
Check by issuing the command
crontab -l (as root)
and also check
/etc/cron.daily, /etc/cron.weekly, /etc/cron.monthly
I can confirm its not an environmental cause. Temperature is very stable in my office.
I’ll give it a try with Glances. Should I follow this procedure to install it on a Raspberry Pi4 ? CrownCloud Wiki - How To Install Glances On Debian 11
Up and running, I’ll keep an eye around 2:00 a.m.
@anon94216691, if Glances doesn’t help, I’ll look with Crontab
@Mikefila, @anon94216691 and @tom_l, pretty sure I already found the cause:
Plex Media Server integration is kicking in every night at 2:00 a.m. I will look into config of that Plex Media Server to see if I can skip the daily scheduled process and rather schedule it once a month.
