How to get the alert if server down.

[kloczek]

You just found missing part in OS Windows template (thx).
Please try to use devel version of this template where I just added this missing trigger with necessary dependencies.

https://raw.githubusercontent.com/kloczek/zabbix-templates/devel/OS%20Windows/OS%20Windows.xml

Please let me know does it work (I have no possibility to test OS Windows template).

[Tharakaw]

Hi kloczek,
The server down will trigger if the Network going down since you are using ICMP and Agent availability to fire the Alert.
{OS Windows:icmpping.count(#10,0,"eq")}>9 and {OS Windows:agent.ping.nodata(15m)}=1 and&#13 I am asking If there any way to alert if server going down with the details(who has shutdown the server, Is it crash, etc)

[kloczek]

Are you sure that you downloaded proper file? You need to download file from devel branch https://raw.githubusercontent.com/kl...%20Windows.xml
In updated template "SYS::Host is down" trigger uses formula:

{OS Windows:icmpping.count(#10,0,"eq")}>9 and {OS Windows:agent.ping.nodata(15m)}=1 and {OS Windows:system.uptime.count(15m)}=0

It raises alarm when there is no new monitoring data for +15m, it is not possible to ping the host.
This trigger is not about that host ":is going down now" it raises alarm trhat host is down indirectly because if host suddenly will be down nothing from inside the system is able to deliver any metrics data pointing that host ius down.

[Tharakaw]

Hi kloczek,
I appreciate your extended reply and the support. As an example let say if the network going down between the Zabbix server and Monitoring windows host(Zabbix Agent) it will trigger server is down since Zabbix Server could not be received from the Zabbix agent.

Thank you.
Best Regards,
Tharaka Wehalla

[kloczek]

If you will be testing at least two independent physical network paths lack of network traffic on one path may allow you to say that it is some network issue.
Do you have some network paths redundancy and/or are you using yet another independent path for only monitoring propose? If not you physical architecture to make necessary diagnostics on top of which will be possible to build necessary metrics and alarming logic.