While setting up an alert for a custom metric it is advisable to configure the alerting period to be either equal to the defined custom metric frequency or the alerting time period to be greater than the defined custom metric frequency for the alerts to trigger properly.
Use case summary :
User create a custom metric group "XYZ" and defined frequency of "15 minutes" . That is, cuegg will be collecting the sample for this custom metric every 15 mins.
In the alert configuration, user has set the notification rule as , if memory utilization exceeds 20% or is utilized less than 2%, for a period of 1 min. (Key point is 1 min)
Based on definition, examiner tries to get the sample data for 1 min, to check the threshold. Since there is no sample on that bucket range (min being 15 mins). There are no alerts generated
Quick solution to that, user should update the freq of custom metrics group to 1 min as well.