Cloudron Forum

I'll have to learn how to look back at the logs. For years everything was fine and stable so I didn't get into the tech much.

For the past month the whole system has been going crazy every few days. The CPU would go to 137% for over an hour and nothing was accessible. The cloudron-support --troubleshoot would slowly show everything was ok, with an added 'fgets: Connection reset by peer'. Everything would eventually come back to life.

I keep hoping an update from an app or Ubuntu would fix it. Next time it happens I'll check top/htop live.

Now updated to 5.11.82

I'm curious if anyone has any insights as to why my invoice ninja would be using an average of 25% CPU for hours, sometimes days, with only one active invoice and practically no user/client activity. It seems to me it should be idle. Thanks for any insight. Thanks
I'm using the current version 5.11.80

@james Thanks for the suggestion. I setup a support ticket with Linode and they responded back that everything looked good hardware wise. They gave me the following suggested commands to check the system:

Check for steal percentages by reviewing the %steal values from the following command:
iostat 1 10

Check for processes stuck in a 'D' state :
for x in seq 1 1 30; do ps -eo state,pid,cmd | grep "^D"; echo "-"; sleep 2; done

Investigate processes using the most CPU time:
top -bn 1 | head -15

Additionally, you can run this command during the excessive CPU usage:
ps fax o uid,%cpu,%mem,time,comm | awk '0+$0 == 999 {print}

Everything checked out normal. I'll keep an eye on it.

@humptydumpty Thanks for the insight. Yeah, it has been pretty stable over the years. And the pdf issue is an easy enough fix when it happens.

I wonder if my payment provider Stripe could be querring my data and causing this. Any ideas on how to check that?

Thanks