Zabbix going insane?

[sporq]

Hello.

So this morning, my zabbix install decided it would repeatedly cycle through all the defined hosts and explain to me that each of them were unreachable. Then they were reachable. Then unreachable. Wash, rinse, repeat.

I'm running version 3.0.12 on RHEL from the yum repos. The install has been online and running for a few years, the update to 3.0.12 happened about a month or so ago and hadn't exhibited any problems before today.

I've restarted the zabbix service and it appears to have calmed down. I have checked zabbix_server log file, and nothing really looked very interesting there. The server is running under ESXi with other hosts, everything hardware related appears to be fine, none of the other hosts on the ESXi server are exhibiting any odd behavior.

Would there be any other logs that may provide insight as to this spazzy behavior?

Regards,

[nelsonab]

My first thought is to check to ensure DNS is working correctly. Next I've seen similar behavior if the poller process is overwhelmed, which can be possible if you recently added a lot of new hosts/items which are all passive (agent default), or the network is operating slower causing the poller work queue to back up.

Lastly Volker's presentatoin from 2015 may be a good source of things to look at: https://www.youtube.com/watch?v=Zwm9qUfH8hU

[sporq]

I shall check that out, thanks.