Repeated false warnings - host unreachable

[cyc]

Hello,

I am having a strange problem. Without any changes in configuration of zabbix (and after running it for several weeks without any problems) I started receiving e-mails about a host being unreachable.
But the host is reachable, and there is another problem - I am getting these e-mails every four minutes. In between, I get messages with the opposite information - server unreachable: OFF.

I tried restarting zabbix_server and zabbix_agentd on the host that's causing problems ... with no success. I am monitoring 10 hosts, but the problem I have is causing only one single host. Or - to be precise, zabbix_agentd on that host, because all the time, the host is perfectly reachable.

[dreas]

I am experiencing the same issue. We are monitoring several servers with Zabbix but for one of the servers I often receive a warning that the server would be unreachable. It is not however. About a minute later I usually get the notification that it is reachable again. This is just happening for 1 of the servers. Any idea why?

[marc]

well i know this issue just in cases the line is buggy or cpu load takes down agent responses. i would check cpu load and availability (ping with reasonable packet size over a long term and check for packet loss).

also it might be good to combine zabbix "status" and icmpping. if both failing, host is down.

at least.. to avoid false positives, customize your trigger expressions and not use last then min(n)=2 instead where n = minimum value for trigger over period of time = 2 (trigger status false) alert me.. so a single negative check want wake you up..

[Libila]

I'm using

{name.template:icmpping.min(300)}=0

I only want the trigger to become true when the ping fails (value of zero) for a total of 5 minutes. (item checks every 60 seconds). Yet with the above trigger it becomes true instantly. Can you paste your trigger marc?

[marc]

I'm using

{name.template:icmpping.min(300)}=0

I only want the trigger to become true when the ping fails (value of zero) for a total of 5 minutes. (item checks every 60 seconds).

i guess {name.template:icmpping.max(300)}=0 is correct for what you want.
"min" would get true on first negative response.

iam using {name.template:icmpping.max(300)}=0& {name.template:status.min(300)}=2 for hosts were agent is available

[bee]

mine:
{name.template:icmpping.sum(#3)}=0

It will fire up action after 3 times unreachable(0).

Thanks,
BEE