Zabbix 5.2.7 - 504 Gateway Timeout with User Role Admin

[bradyjg]

Problem:
I'm having a problem with Zabbix 5.2.7. I have an account let's say user "A", which is part of a default Admin user role and a user group with read-write permissions for multiple host groups.

When I go to the dashboard for any of the host groups while logged in as user "A", The Trigger Overview and Problem widgets continuously spin and give me a HTTP 504 Gateway Timeout in Chrome dev tools. This also happens when accessing the Problems page under monitoring.

On the other hand, when I go to the dashboard as a Super Admin, user "B", I can see everything perfectly fine with no 504 response and it is almost instant when loading, so I know that it isn't too much data and is actually timing out.

Attempted Solutions:
Checked the log files for server and agent - no errors relating to this.
Created a new user, user role, and user group with correct permissions - same result.
Tried creating a fresh dashboard with a new widget for problems - same result.

Configurations:
I have attached the configurations for the widgets.

The devs said this was not a bug, so they cannot help me. Can anyone help me see what I am doing wrong?

[Glencoe]

Is this Apache? Are you sure you checked all *web server* (Apache or Nginx) logs? Many Apache configs split access.log and error.log, and error.log would be the best place to check for such problems.

[bradyjg]

Thanks for the response. I am using nginx and have checked both access and error logs. We found the issue though.

Since we are using the InnoDB cluster, we had to remove all the foreign keys when deploying the database and did not put the indexes back, therefore creating queries long enough to timeout our HAProxy. So our issue was on the database side.

Again, thanks for taking your time to help.

[damiancancela]

Hello, im having same issue with 5.0.23. Were you able to fix it? I thought upgrading to latest build will help but did not.
Thank you

[bradyjg]

Yes, it had to do with an unrelated issue with Zabbix. Our HAProxy instances were timing out because we did not index our database correctly.