Server "..." is unreachable for 3 minutes - Alert issue

[robcc]

Hi there,

I'm realy new to both Zabbix and Linux so please bare with me,
I have managed to get zabbix 1.8.3 up and running on the CentOS platform, I only have 9 hosts monitored with a total of 378 items being monitored.
I have configured zabbix to send an email to myself when a host is unreachable for 3 minutes.
The problem I am having is that i am recieving emails every hour or so from random hosts being unavailable when they are still working correctly and i am able to RDP into them.
This is a big issue as I am now starting to disregard the emails I am recieving which means I miss when a host actually goes down.
I don't think the zabbix server is overloaded as i am only monitoring 9 hosts.

Please Help?

[MrKen]

Might be useful if you post the trigger expression that you are using.

MrKen

[robcc]

The Trigger Expression is "{HOSTNAME:agent.ping.nodata(3m)}=1"
I don't think it's a problem with the trigger because when it reports that the host is unavailable, no other data such as CPU usage is collected

[MrKen]

I hope your trigger expression actually says: nodata(180)}, as it should be in seconds.

There could be any number of reasons why Zabbix reports a server as being unreachable, but seeing as you are on 1.8.3, you might be experiencing this bug, which has now been fixed in 1.8.4

Recommend
1. Check trigger expressions.
2. Upgrade to 1.8.4
3. Use icmpping instead of, or together with agent.ping

MrKen

[robcc]

I have not changed any of the triggers they were set up from the template_windows template.
I will change the trigger to seconds just in case.
But it's probably worth upgrading to 1.8.4 just to fix any of the bugs isnt it.
Is the upgrade easy to perform?

[MrKen]

Upgrading form 1.8.3 to 1.8.4 should be a breeze.

http://www.zabbix.com/documentation/1.8/manual/installation/upgrading#upgrade_procedure