Monitoring Host - Notification after X time

[pat_trick]

Hello,

I've been enjoying working with Zabbix and we're currently importing all of our checks from another tool. One thing I'm not sure how to do is delayed timeouts for a notification.

Here are the properties of an example check for a host:

TCPConnectionCheck	Triggers (1)	net.tcp.service[tcp,,80]	30	90	365	Simple check

My understanding is that this does a TCP service check against port 80 on the host every 30 seconds.

Next, the Trigger is as follows:

{<hostname>:net.tcp.service[tcp,,80].last(0)}=0

Which works great whenever there's an issue with the TCP to port 80. However, what I want to do now is make it so that the Trigger doesn't fire until the issue is occurring for at least 10 minutes. If it resolves in that time frame, I'm not worried about it. I think that this would require use of the count() function, but I don't know how to write out the Trigger to appropriately check on this; the examples I've seen on the forums don't seem to be specifically applicable.

Any assistance is greatly appreciated.

[heaje]

Here's how I would do it:

{<hostname>:net.tcp.service[tcp,,80].max(10m)}=0

That trigger looks at all values in the last 10 minutes. If the maximum of all those values is "0", then the port has been down for 10 minutes. Another variation is:

{<hostname>:net.tcp.service[tcp,,80].max(#3)}=0

This one looks at the last three collected values. If the maximum for all of them is "0", then the port has been down too long.

[heaje]

A good place to look for information on the available trigger functions is at https://www.zabbix.com/documentation...gers/functions.

[pat_trick]

Thank you very much; this is the exact information that I needed. It doesn't spell out in the function parameters that you can count in minutes.

Now to just work out these dependencies!