anyone experiencing performance problems with the active agents?
release 2.2.1
we have around 350 hosts, each with 30 monitored items.
total performance about 37 values per second.
everything is fine when all the items are passive, but:
as soon as i change just 2 items to active (update interval each 8 and 12 hours only!),
"Zabbix agent (active)" queue starts growing immediately.
with 1 active agent check the "More than 10 minutes" column is around 30 items.
StartPollers=65
StartIPMIPollers=2
StartPollersUnreachable=40
StartTrappers=65
StartPingers=15
CacheSize=64M
StartDBSyncers=10
Timeout=30
ValueCacheSize=32M
any idea what could be wrong?
//EDIT:
it seems that not only the active queue starts growing, but also the "Simple check" and "Zabbix agent".
this is weird. with all passive items there is almost no queue, but with 1-2 active checks there are 3 queues.
release 2.2.1
we have around 350 hosts, each with 30 monitored items.
total performance about 37 values per second.
everything is fine when all the items are passive, but:
as soon as i change just 2 items to active (update interval each 8 and 12 hours only!),
"Zabbix agent (active)" queue starts growing immediately.
with 1 active agent check the "More than 10 minutes" column is around 30 items.
StartPollers=65
StartIPMIPollers=2
StartPollersUnreachable=40
StartTrappers=65
StartPingers=15
CacheSize=64M
StartDBSyncers=10
Timeout=30
ValueCacheSize=32M
any idea what could be wrong?
//EDIT:
it seems that not only the active queue starts growing, but also the "Simple check" and "Zabbix agent".
this is weird. with all passive items there is almost no queue, but with 1-2 active checks there are 3 queues.
Comment