No incident being created when alert condition goes into critical violation

I created a simple alert condition based on NRQL selecting a custom event that I created via the API. I just want the alert to create a violation if there are no events within 15 minutes. I have set the query threshold and data aggregate window so that it shows correctly in the graph when there were critical violations. The problem is no violation actually is getting created.

Hi, @joeacool: If no events are present, count(*) returns null, not zero, so it will never violate your alert condition. You may change your query to SELECT count(attr OR 0), or configure your condition to violate on loss of signal.

I’ve made both the change for loss of signal and to the condition. Still, no incident is being created. Maybe I’m missing something. As configured if there haven’t been any of these custom events for 15 minutes an incident should be created right?

@philweber Can you look at this configuration? I tried a brand new account with the same type of alert. Again, the graph shows the incidents, but they are never actually created by the alert.

Sorry, I think you will need help from a support engineer (I am a trainer). Someone will help you here as soon as possible.

I’ve seen multiple threads saying this same thing is happening, is there something account specific? None of the people with the same issue seem to be getting any answers from engineers.

Hi ~
Thanks for hangin’ in there - let’s see what else we can learn and how we can help

Can you share a permalink to your monitor (don’t worry, it’s only viewable internally here at New Relic) so that we can take a look at your monitor and more accurately provide feedback?

I know you’ve been combing the forums, but I also wanted to highlight some docs around the alert and incident workflow:

:arrow_right: A walk through of the Alert creation workflow

:arrow_right: Identifying/managing when an issue is created

Let us know if those resources help, if not please share a permalink to your monitor and we’ll take a look.

I’ve walked through the documentation, to me it looks configured right. Again you can see today I pushed events so that the it went out of critical threshold, and then when it goes back in no incident is created.

Actually when I triggered it today it is working. I didn’t change anything in terms of how the alert or condition was set up, so I’m assuming something was fixed with incident generation, thanks!

Great to hear that! Thanks for letting us know