Netdata updater saturating RAM and CPU

HI all,

I have had an issue on my server which took me quite a long time to find.

Since multiple month, I have a RAM/CPU usage issue on my OVH server every morning, starting at 6:30 am, Paris time.

After a lot of search, I found out that it’s coming from the daily crontab scripts, and more precisely the Netdata-updater script.

I have stopped the daily crontab that launch Netdata updater for the moment, could you please explain how to solve this issue ?

It’s my first discourse post so I don’t know yet how to add picture to the post, but I have all the data/charts needed.


Regards

Hi there,

Anyone to get me a hand on this one ?

Regards

Hello @FloDus and welcome into our community, could you provide me with your netdata -W buildinfo and the platform of your system?

Hi Tasos,

I have been using Netdata on an OVH server running DEBIAN.

Here is the result of the cmd :
Version: netdata v1.33.1-86-nightly
Configure options: ‘–prefix=/usr’ ‘–sysconfdir=/etc’ ‘–localstatedir=/var’ ‘–libexecdir=/usr/libexec’ ‘–libdir=/usr/lib’ ‘–with-zlib’ ‘–with-math’ ‘–with-user=netdata’ ‘–with-bundled-protobuf’ ‘CFLAGS=-O2’ ‘LDFLAGS=’
Install type: unknown
Features:
dbengine: YES
Native HTTPS: YES
Netdata Cloud: YES
ACLK Next Generation: YES
ACLK-NG New Cloud Protocol: YES
ACLK Legacy: NO
TLS Host Verification: YES
Machine Learning: YES
Stream Compression: NO
Libraries:
protobuf: YES (bundled)
jemalloc: NO
JSON-C: YES
libcap: NO
libcrypto: YES
libm: YES
tcalloc: NO
zlib: YES
Plugins:
apps: YES
cgroup Network Tracking: YES
CUPS: NO
EBPF: YES
IPMI: NO
NFACCT: NO
perf: YES
slabinfo: YES
Xen: NO
Xen VBD Error Tracking: NO
Exporters:
AWS Kinesis: NO
GCP PubSub: NO
MongoDB: NO
Prometheus Remote Write: NO

Regards

What instructions did you follow to install the Netdata Agent?

DEBIAN.

Specific version?

Resuming to this, to give you a solution before the end of the day.

What instructions did you follow to install the Netdata Agent?

I just want to understand how you end up with an Install type: unknown

An action you could take to overcome your issue is:

Try to run the kickstart script (Install Netdata with kickstart.sh | Learn Netdata) again with the option --reinstall-even-if-unsafe. This will reinstall the Agent. The kickstart script will install the Agent in 3 possible ways (in order)

  1. Native packages (if we support your system Netdata platform support policy | Learn Netdata, in your case Debian. Debian 9/10/11 should would just work)
  2. Static binaries
  3. Local build

The reason behind your problem (a speculation until you provide me with an answer how you installed the Agent) is that in every update the agent recompiles it’s components (local build). If you install the Agent with the 2 first methods you would not even notice any activity between updates.

Since you’re on a Debian system, I suggest you do a full uninstall and then install again with kickstart.sh. It will default to use our binary packages, which will eliminate the CPU/Mem spike. If you have user configurations under /etc/netdata, keep those files somewhere before you uninstall, so you can add them again after the new installation.

We’re a bit cautious to not mess up custom installations, so the previous installation type is used to guide kickstart in deciding whether we can use binary packages or not. In your case we can’t.