Utilization of timer processes over 75%

[emackey3]

Zabbix-server version 7.09
Database is on another server postgresql version 13
We are monitoring 1204 hosts.

We have recently started getting Utilization of timer processes over 75%. The problem is intermittent with most of the errors happening overnight.

When the issue started the StartTimers in zabbix_server.conf was 10. I've slowly increased this to 30 but it hasn't resolved the issue.

Both the application server and database server aren't under any cpu or memory load during these times.

I see a connection timeout in the database logs that correspond to when we receive the alerts. "LOG: could not receive data from client: Connection timed out." A dba has confirmed that the database is not running out or connections and seems to be in good order.

We have reached out the network team but nothing has been found.

I didn't find anything useful in the zabbix application log.

Anything I can do to additionally troubleshoot this issue?

[rkollman]

The timer process becomes active when you have maintenance periods running. So as you write the errors with timer process happens overnight, I think you have a maintenance periode running. The Timer process checks for problems which arise during this period and closes them for hosts in maintenance.
So just check your maintenance periods and if you have a lot of them (or a lot of hosts in maintenance), increase the timer processes more than you already did.

[emackey3]

Thanks. We do have lots of maintenance periods and hosts in those maintenance periods. I've upped the start timers to 60, hopefully that resolves the issue.