Ticket lost in memory usage peak

[Palbin]

Infos:

• Used Zammad version: 3.2.0
• Used Zammad installation source: source
• Operating system: Ubuntu 18.04.1 LTS
• Browser + version: Chrome

Hello,

We are having some problems with Zammad memory usage. The following graph shows the memory behaviour:

When the system reached the first plateau, we stopped receiving new tickets until we reboot the system. To avoid this, we rebooted the system before it reached the peak the second time (on June 12th).

We have set WEB_CONCURRENCY to 4 and we don’t use elasticsearch. We also have set this variables (from this thread Memory usage in Zammad):

export RUBY_GC_MALLOC_LIMIT=1077216 RUBY_GC_MALLOC_LIMIT_MAX=2177216 RUBY_GC_OLDMALLOC_LIMIT=3000100 RUBY_GC_OLDMALLOC_LIMIT_MAX=4000100

Our main problem is the ticket lost in the memory peak, is there any solution for this problem? It could be a setup issue?

Addicionally, any tips or hints to improve the performance would be appreciated.

Best regards.

[MrGeneration]

Well… how do you determine that you’re having a ticket loss?
Please describe your issue - whar are you doing and what’s happening.

Right now I don’t know how you’re determining your data loss which makes it impossible to help you.
Please also have a look into your production log which might also hold information.