Zammad consuming all RAMs

Technical assistance
1

nfos:

  • Used Zammad version: 6.0
  • Used Zammad installation type: (source, package, docker-compose, …) package
  • Operating system: Debian 10.10
  • Browser + version:
    Chrome, Edge, Firefox, private windows, to no avail

Expected behavior:

  • We had RAM issues with elasticsearch then we added it fro 16GB to 32GB, after only a few weeks, now it seems Zammad is using most of the RAM

Actual behavior:

  • The web UI will hang from time to time

Steps to reproduce the behavior:

Check the zammad resources via top command, only around 1GB free RAM and the system behaves abnormally.

zammad memory out
zammad memory out874×586 148 KB

zammad hanging

2

Have a look at the mem persentages. It’s not a ruby process but a Java process taking 50% of your memory. If I’d had to guess it’s javascript.

Look up their documentation for “heap size”. You can limit the memory consumption of Elasticsearch. 2-4 GB memory should be good enough.

Also, press “c” in your top to actually see which process exactly is running that will help you in future situations. Or use htop for easier sorting CPU vs Memory etc etc.

That’s a different unrelated issue.

3

Dear Mr,

Would you advise how can we limit the memory consumption of Elasticsearch in Zammad server side? I’ve increased the RAM to 32GB and I guess if we limit Elasticsearch to maybe 4GB then the rest will be used as caching purpose, this is even better for speeding up the searching by the end users and agent, is my understanding correct?