huge SNMPv2 queue

[herta]

Environment:
- zabbix server
OS: Ubuntu 12.04.3 LTS
Zabbix server: Zabbix server v2.0.4 (revision 31984) (08 December 2012)
MySQL: mysql Ver 14.14 Distrib 5.5.34, for debian-linux-gnu (x86_64) using readline 6.2

Since Saturday, zabbix has been bombarding us with "Zabbix agent on xxx is unreachable for 5 minutes" mails. I've been tuning pollers, caches and innodb. The zabbix server graphs all look OK now, but the queue remains extremely high. What is odd is that it are mainly SNMPv2 agent items that don't seem to get processed (over 4700 items in "More than 10 minutes").

Does anyone have any idea what might be going on or how I can debug this?

Kind regards,

Herta

[tchjts1]

See this post as another member had that same issue and found a solution: https://www.zabbix.com/forum/showthread.php?t=38804

[herta]

Thanks for your reply, tchjts1. I had read that post (and a number of others), but as I wrote: there are no errors in our logs.

In the meantime my colleague found the problem. One of our Cisco switches had +80 items with +55,000 triggers. Not sure what caused that. Zabbix couldn't open the list of triggers to investigate further, so we ended up unlinking and clearing the template. After that, everything quietened down very quickly.

[Kaykim]

I also encountered the same situation.
and one of my network device has more than 2500+ items 670+triggers 230+graphs .....
there are huge queue on SNMP Agent....