Zammad railsserver performance

Technical assistance
1

I am running zammad on a docker container with the following host specs

RAM: 126GB
DISK: 1.8T

Memory usage for the railserver continuously increases, even though no users are using it

What environmental parameters I set for railserver

   - WEB_CONCURRENCY=8
   - MAX_THREADS=8
   - MIN_THREADS=4
   - ZAMMAD_SESSION_JOBS_CONCURRENT=8
   - ZAMMAD_WEB_CONCURRENCY=8
   - MALLOC_ARENA_MAX=2
   - RUBY_GC_MALLOC_LIMIT=1077216
   - RUBY_GC_MALLOC_LIMIT_MAX=2177216
   - RUBY_GC_OLDMALLOC_LIMIT=3000100
   - RUBY_GC_OLDMALLOC_LIMIT_MAX=4000100

Screenshot 2024-01-10 at 10.00.45
Screenshot 2024-01-10 at 10.00.45842×215 10.6 KB

it will increase until the RAM overflows, someone has encountered this problem and solved it before please help me about it

2

What version are you talking about?

3

Thank you for your feedback

we using version zammad -5.0.2-1

4

Your version is terribly outdated. Zammad 6.2 is current. Please upgrade.

5

Does that solve our problem of server performance increasing when starting Zammad @MrGeneration

6

I just want to make sure it can help us, and are there any performance changes, and what do I need to install to make it better?

7

You’re using the term “performance issue” but the only thing you actually complain about is the memory usage. Your Zammad version is over 12 months old and has security advisories.

Patching in general helps to get rid of bugs, keeping it up to date etc etc.
If that’s not enough arguments for you, then I guess, no, you don’t have to update.
Zammad is not a set and forget piece of software.

8

Sorry if I’m using the wrong terminology here, but this affects performance, increased CPU causes all operations to stall, and this happens when no one is using it, it always takes up memory, I wondering if upgrading will make it faster and without this error

image
image768×288 13.9 KB

9

I created elasticsearch index, and used 32GB RAM for it, and I tried to find solution to solve it but no expected effect

10

HTTP/503 is rendered by your proxy server.
The log file of that proxy server will tell you why it threw the error and that’s the part that will help you to find the issue you actually have.

That being said, you should upgrade to a current version (6.2 as of now) to cross out that this is a potential bug underneath on top.