we recently had some issues (slow behaviour) on git.centos.org and it seems related to mod_wsgi settings mostly using default values.
We have tuned the fedora side some, but I would be interested to hear how this turns out for you. ;)
https://pagure.io/fedora-infra/ansible/blob/main/f/roles/pagure/templates/0_pagure.conf#_6
@kevin : this is our default one but what I'll have to tune (but analyzing mod_wsgi options) : https://github.com/CentOS/ansible-role-pagure/blob/master/templates/httpd-pagure.conf.j2#L6
Clearly bumping number of processes to match cores number on the host and trying other options from https://modwsgi.readthedocs.io/en/master/configuration-directives/WSGIDaemonProcess.html?highlight=listen-backlog (as we hit that limit today)
Metadata Update from @arrfab: - Issue assigned to arrfab
Metadata Update from @arrfab: - Issue tagged with: centos-common-infra, high-gain, investigation, medium-trouble
I read some settings and processes/threads seem to be the common one so just declared as dynamic variable (based on ansible fact and number of vcpus/cores but can be changed)
So it can also now be statically forced through ansible inventory at the group_vars/host_vars level instead of default values (keeping an eye through zabbix monitoring on the effect)
Metadata Update from @arrfab: - Issue close_status updated to: Fixed - Issue status updated to: Closed (was: Open)
Log in to comment on this ticket.