I'm trying to solve problem with queue of delayed items data.

After few hours after start zabbix I'm able to observe that telnet on server port does not open instantly TCP session. After this all active checks have longer and longer delay. Sometimes it recovers but sometimes not and so far only restart of server helps.

Looking on netstat output on zabbix server I see:

# netstat -an | grep <IP.of.zbx.srv>:10051| awk '{ print $6}' |sort| uniq -c
38 CLOSE_WAIT
99 ESTABLISHED
26 SYN_RECV
1 SYN_SENT
396 TIME_WAIT

so numbers are not big.

Because it was not possible to open new connection to zabbix srv I've increased StartTrappers to 40. After this after few minutes I had tons of error messages in server log like:

31964:20131013:143013.211 [Z3005] query failed: [1205] Lock wait timeout exceeded; try restarting transaction [update ids set nextid=nextid+1 where nodeid=0 and table_name='events' and field_name='eventid']
zabbix_server [31964]: ERROR [file:db.c,line:1582] Something impossible has just happened.

With StartTrappers=40 zabbix server is almost instantly unuseable.