Netdata service freezes VPS completely


Installing netdata on clean Debian 11 VPS causes server freeze.

I have a VPS (1 core 2.2 GHz, 1GB RAM, 20 GB SSD, Hyper-V) at and after installing netdata I get server freeze while netdata service is running. I had experience with running netdata on other nodes with 1 GB RAM and that was quite ok. It takes about 10-15 minutes to type command ‘sudo service netdata stop’ to make VPS responding again.

/proc/sys/vm/max_map_count is set to 2000000
swap is 2 GB
swappiness is set to 15

When netdata is up, CPU utilization becomes 100%. Sometimes top reports on 50% or more CPU utilization by kswap and 10-15% CPU by netdata, sometimes it reports as 97-98% wa (iowait) and other processes less than 1%. RAM allocation becomes almost 100% too. Stopping netdata service releases half of RAM utilization and CPU utilization becomes around 2-4%.

Screenshot of top output with netdata runnung:

Screenshot of top output with netdata stopped:

Logs are here - netdata error.log, access.log and netdata output

What I expected to happen

Netdata is running on the node as usual without freezing the server.

Hi @alpham !

Ok, obviously Netdata is eating most of the ram on your system. We can try to slim it down perhaps using netdata.conf. Are you running a stock configuration right now? Are there any plugins that you might not use and we can turn off? Could you try perhaps instead of dbengine another [db] mode (e.g. ram) ?

I will also try to run it in a 1GB environment and see how it currently fares.

Hi, @Manolis_Vasilakis

I suppose that is hosting-related situation. On other nodes with almost same config (Ultahost, DO) I never encountered such problem.
Yes, I’m running stock netdata.conf with only difference that this node is attached to my Netdata Cloud and changing [web] mode = none. On my VPS platform the situation reproduces every time I reset VPS to fresh Debian 11 installation and then run one-command netdata install script (however last time it took about 10-14 hours after the reset when situation reappeared)

As this node is not a production system anymore, I can provide sudoer’s login/pass to Netdata staff member if it would help track the problem. If that would help, please PM me and I’ll provide my email (or you can just email me from any address if you can have my email from my forum profile)

Also I forgot to mention that after fresh install of netdata I get this error when run netdata’s edit-config:

/etc/netdata/edit-config: 26: /usr/share/bash-completion/bash_completion: Syntax error: "(" unexpected

I don’t get this error when I comment out 2nd line in /etc/netdata/edit-config:
#[ -f /etc/profile ] && . /etc/profile

I can get back to the node on Wednesday and try to change config to reduce memory consumption.