Can we create a alert to trigger when given threshold value crosses it threshold value and stays there for 30minutes ?
If my threshold is 100 value and measure reaches that value and stays constantly >100 for 30minutes, I would like to create an alert.
Solved! Go to Solution.
I believe you could achieve this by changing the aggregation of the incident you are configuring to max.
This way, the maximum value would have to equal or exceed your threshold for 30 minutes before the alert fires.
Let me know if that works for you or if you need further assistance.
I think you have the right thinking, but if you used the max aggregation it would actually fire as soon as it got one return value above the threshold. Max says take the highest value in that timeframe, so if you saw results 80, 75, 67, 101, 89 over a period of time it would take the highest value and immediately fire.
The way you can get behavior like this for alerts in your case would be to add the measure you're interested in to the incident and since in your case you want to alert if the value is above your threshold for 30 minutes set the aggregation for the condition measure to 'minimum.' Then set the evaluation timeframe to 30 minutes.
Using the minimum aggregation says to take the lowest value seen for the measurement at each interval and in combination with the 30 minute timeframe you'll be getting the lowest value recorded over that period as the number that is evaluated. If you think about it this means that in order for the incident to be triggered, you'll have had to have been seeing measurements above your threshold for that entire 30 minute window. You can adjust the window to get different levels of 'extended' outages.
Note that if you had a measure where a higher number is better than a lower number (think content verified or host reachable for a url monitor where 1 is good and 0 is bad) you would select the maximum aggregation to watch for extended issues.