Queue of items to be updated

[markymark33]

Hi

I know this has been discussed before, but I have a constantly large queue count of over 8000 and I cannot resolve the issue. The alerting starts getting delayed and the delay increases to over an hour.

I've tried the tweaks for the server conf file mentioned in other threads but to no avail.

The hardware is the latest nehalem based servers for both server and db.

I have ~500 monitored hosts with a count of 260 nvps.

Can someone first of all explain what the queue count is ?
Is it items that are to be written to the DB or processed by the zabbix server.



Any help will be much appreciated.


Thanks



Mark

[GArmao]

just a question, have you enabled Distributed Monitoring on the server config (NodeID parameter != 0)?

[markymark33]

Its a standalone setup so NodeID = 0

Version is 1.8.2

[MrKen]

There are many reasons why you could have a long queue.

Database tuning could be one.

Another could be dns problems. Is /etc/resolve.conf properly configered? Can you get a result from :# dig @name_server <randomhost_ip_or_dns> ?

Also try:# /etc/init.d/nscd status (If not running, start it).

HTH
MrKen

[bashman]

I recommend you to upgrade to v1.8.3 as soon as you can, there are performance improvements.

Do you have Zabbix frontend Web server, Zabbix server and db on the same host?. Could also be an IO problem.

You can do Web server and db tuning, also you can tune zabbix frontend (refresh intervals, rows, autologout, etc).

Try to keep nvps low by increasing your polling interval.

I think the zabbix queue gives you the item delayed time which was supposed to be updated.