about trigger dependency

[ykdhan]

Let's say I have two hosts. One is a firewall and the other one a L2 device.
I set ping check items with different time interval. 1-minute interval for firewall, and 5-minute interval for L2.
And I also set triggers for both of them just to be informed when there is a disconnection or error. And the FW trigger is dependent on L2 trigger.
If there is a problem in FW, I should get a notification of that trigger, and if there is a problem in L2, I should get one notification from L2 trigger only.
However, I am just curious how this works even though their ping check time intervals are not the same.
Regardless the time interval difference, does Zabbix check L2's ping right after it detects any problem in FW so that it determines whether the problem is from L2 or not?

[tim.mooney]

It does not.

Zabbix trigger dependencies are great, but the idea you're talking about is one area where I believe there is useful functionality that's missing.

What we've done to reduce (but not completely eliminate) "false positives" is to set the monitoring frequency for the network topology dependencies to be more frequent than the other triggers that depend upon the network topology triggers. That way, we're more likely to detect a problem with the underlying dependency, rather than seeing the problem "first" with a host or bunch of hosts.

[ykdhan]

Thanks for your reply. Well, is there really no way to differentiate the ping check interval to find the real origin host of a problem? I have about 7000 firewall devices to monitor and I do not want to add another 7000 L2 devices with frequent check.