items not being checked systematically

[edeus]

Greetings

I have an odd problem with 4 hosts, agent.ping would just stop being checked randomly. agent.ping is checked every 60 seconds.

When I look at the zabbix queue it seems to have a lot of items randomly across the board.

1) If Zabbix is running effectively, should there be waiting items in the queue or should it be empty.
2) Does using the zabbix items: zabbix[*] i.e. zabbix[history], zabbix[queue], etc caused any performance issues. I just realised it may be running select count(*) on the history table every 30 minutes which would definitely cause slow down.

There are no odd errors in the log files about agent.ping on these 4 specific hosts. The other servers are monitoring agent.ping fine. It is always these 4 specific hosts that zabbix doesnt send out agent.ping's for.

Verbose debug shows no agent.ping requests to these hosts either so its something on the server side. (Agent versions are also updated to 1.1 or 1.1.1 and have virtually identical configs)

Thoughts, opinions and ramblings would be appreciated!

[bashman]

Why are agent.pings getting stock in the queue?. Other items from the same hosts don't get stock, only happens with agent.ping.

The problem is that I get plenty of false mails about agent unavailability due to the nodata() function, but agent is responding to the rest of items.