Bad performance and mail / scheduler failures

Infos:

  • Used Zammad version: 6.2.0-1707115064.33614ef9.focal
  • Used Zammad installation type: Package
  • Operating system: Ubuntu 20.04 on a VMware
  • Browser + version: Chrome 121.0.6167.185
    Hardware Specs: 12 Core - 24GB RAM

Expected behavior:

Normal running System like outside working hours

Actual behavior:

Everyday when all Agents logged in (about 80 Users) we get a bad performance and failures in the Monitoring

scheduler may not run (last execution of SessionTimeoutJob.perform_now about 2 hours ago) - please contact your system administrator
Channel: Email::Account in is active but not fetched for about 1 hour
Channel: Email::Account in is active but not fetched for about 1 hour
Channel: Email::Account in is active but not fetched for about 1 hour
Channel: Email::Account in is active but not fetched for about 1 hour
Channel: Email::Account in is active but not fetched for about 1 hour
Channel: Email::Account in is active but not fetched for about 1 hour
Channel: Email::Account in is active but not fetched for about 1 hour
Channel: Email::Account in is active but not fetched for about 1 hour
Channel: Email::Account in is active but not fetched for about 1 hour

We have 9 Mail Accounts for 12 Groups

Steps to reproduce the behavior:

Activate the maintenance mode and restart the zammad-worker. After a few minutes it will works fine for a few hours and then the failures came back.

I Already set the ZAMMAD_SESSION_JOBS_CONCURRENT to 4.

Is there any chance to fix that problem?

Thanks

You’ll need to fiddle around to find a proper combination of performance tuning sets.
That’s always a big complicated and fiddling. Session jobs concurrent definitely is not enough.