APM Variables Not Working in Advance Alert???

hi there

is the trigger running in the application component context? (using the pull down on the trigger tab)

cheers

dan

Can you please open a ticket on this one and post your case number here?   I'd like to have one of our support reps take your case and ensure that you get up and running.  My personal motivation for doing this is to ensure we can make this process more intuitive moving forward ;-)

What was the solution for this, if you don't mind sharing.  I just upgraded to 4.0.1 over teh weekend, and I'm seeing the same problem.

Dwyane

Hello,
If you are experiencing this problem just after the upgrade please go ahead and open a support ticket. It would be just guessing without knowing the context.

Otherwise I would check if macros are evaluated in relevant context (as mentioned above). I mean ${ApplicationName} will work when monitored property is "APM: Application" or "APM: Component" but not on "Node" etc.

I can't speak for the above, but my issue appears to be something else.  Here's what I found...

We used the unmodified (accept adding e-mail notification) "APM: Component" alert that comes with APM 4.0.1, and tweeked the threshold of an application component to trigger a critical condition.  The application monitor contains several components, and we tweeked a component in the middle of the list.  When we ran "Test," we received all kinds of variable names and misleading data in the e-mail like above, and found that the "Test" was using the details from the first component in the list and not the component in a critical status.

However, when we enabled the alert, it properly triggered and properly formatted the notification.

This appears to be a bug in the advanced alert test function.  I'd be happy to get with support and troubleshoot further, if you think it is necessary.

Dwyane

Testing alert is quite different than triggering. The purpose of the test is to verify that all trigger actions will be executed.

In test trigger conditions are (partially) simulated and should not necessarily match actual state of net objects. This also means test could be triggered on objects that are not matching trigger conditions and all actions should be executed.

For instance I can run test alert that is fired when group goes down even though I have no group at all.

I would say this is by design. Like you want to make sure your SMTP setting is true or file you want to execute is on the place and don't want to plug out cable just to get node down.