Issue with the queue and a single host

[Whitestarb5]

I am having difficulty with the queue length and a single node.
The server is not in heavy use items, new values per second is 36.
The server is on dedicated hardware and the database does not appear to be busy; show processlist shows nothing queuing,
Busy pollers etc remain under 5% at worst.

Whilst items occationally appear in the queue for all hosts one in particular appears to have gotten completely stuck. Some values have not been updated for hours but others are up to date.

This server does perform the most user parameter based checks but the logs are not showing failures.

I do see items like this in the log
Zabbix agent item [NAME[PARAMETES]] on host [lampext01] failed: another network error, wait for 15 seconds
They are always followed by
resuming Zabbix agent checks on host [lampext01]: connection restored

Pings to the host from the zabbix server seem fine and zabbix_get return the expected data.

I am running out of ideas on how to further optimise what is a fairly small Zabbix instance.

Any thoughts gratefully received.

Richard
Zabbix v1.8.11

[Whitestarb5]

Timeout Values on Server and Agent Config seem to have helped

I increased the TimeOut values on the Server and worst affected agents from 3 to 10 and this seems to have helped.

Some of the UserParameters may have been running slowly (relatively) and locking up agents, not sure but it seems possible.