The system runs for an extended period of time with nominal CPU usage.
Actual behavior:
After a few days of running several cpu-heavy tasks keep on running for hundreds of hours at a time, thus leading to a critical cpu load.
My environment variables are set as following:
WEB_CONCURRENCY=5
ZAMMAD_SESSION_JOBS_CONCURRENT=8
ZAMMAD_PROCESS_SCHEDULED_JOBS_WORKERS=1
I have attached a screenshot of the current most cpu-intensive tasks executed on the server. I hope this helps.
Steps to reproduce the behavior:
Keep the system running for an extended amount of time.
Thanks for your swift answer! What I mainly was wondering about was why the cpu usage is drastically lower and only starts reaching very high loads after some days. Could it be that the allocated ressources are just not enough for the performance tuning in place (leading to job backlog etc…).
I’m currently running an instance with around 20 active agents and around 500 active tickets. Are my environment variables overkill then and if so what settings would you recommend?
I’d say it’s overkill for 20 agents, yes, however, it highly depends on your instance and its configuration. And of course the usage.
If you have huge tickets with a lot of content, web concurrency might have to be higher for more throughput data wise.
If you only have 20 concurrent agents and require 8 session workers, then I’d say this is definitely out of the ordinary. This is something I’d set for 100 agents plus (most likely rather 200 plus).
