skip
April 3, 2024, 10:53am
2
Question out of curiosity, what does your server infrastructure look like?
And what means 1000 agents? Concurrent agent sessions at the same time?
ajitdalvi:
version: 6.0
Zammad 6.0 is outdated. We’re at Zammad 6.2 which is a security release.
ajitdalvi:
Zammad not working when there are 1000 agents
Instructions unclear. Please provide tuning settings if applicable. And note that it’s about concurrent agents not the total agents you have in your system.
Also tell us “What does not work” that would help a lot.
1 Like
Sorry My bad, We are using version 6.2
zammad config:get WEB_CONCURRENCY =50
zammad config:get ZAMMAD_SESSION_JOBS_CONCURRENT = set
zammad config:get ZAMMAD_PROCESS_SCHEDULED_JOBS_WORKERS = 1
zammad config:get ZAMMAD_PROCESS_DELAYED_JOBS_WORKERS =4
Machine size GCP c3 std with 44 CPU and 176 GB Memory 128 gb standard pd DIsk
Following issue-
Frequent upstream error
DIsk getting full for no reason automatically it comes to normal size
Email Fetching issue with following errors-
Channel: Google::Account in is active but not fetched for about 8 hours
unprocessable mails: 103
scheduler may not run (last execution of CacheClearJob.perform_now 10 days ago) - please contact your system administrator
Search related issues -Search doesnt happen sometime- We have to reindex then it starts working.
Concurrent agent is between 300 to 500
Thanks for reply.
Machine size GCP c3 std with 44 CPU and 176 GB Memory 128 gb standard pd DIsk
Updated details:
Zammad version 6.2
Infra:
Machine size GCP c3 std with 44 CPU and 176 GB Memory 128 gb standard pd DIsk
Tunning:
zammad config:get WEB_CONCURRENCY =50
zammad config:get ZAMMAD_SESSION_JOBS_CONCURRENT = set
zammad config:get ZAMMAD_PROCESS_SCHEDULED_JOBS_WORKERS = 1
zammad config:get ZAMMAD_PROCESS_DELAYED_JOBS_WORKERS =4
Following issues-
Frequent upstream error
DIsk getting full for no reason automatically it comes to normal size
Email Fetching issue with following errors-
Channel: Google::Account in is active but not fetched for about 8 hours
unprocessable mails: 103
scheduler may not run (last execution of CacheClearJob.perform_now 10 days ago) - please contact your system administrator
Search related issues -Search doesnt happen sometime- We have to reindex then it starts working.
vic
April 8, 2024, 1:47pm
10
You’re setting up Zammad for up to 1000 (!) agents on cloud hardware that costs no less than $15k per year, and yet you have not considered getting commercial support from Zammad directly?
3 Likes