I have recently started working with Windows Azure and have read that they have an option of setting up alerts. I tried to set up an alert and found that there was an option called period which is defined as "Time span during which to monitor the metric data specified by the alert rule". This definition is somewhat unclear to me.
For continuous metric like CPU : If I select the period to be '5 min', does that mean that Azure will monitor the metric for 5 minutes and if the metric is above the threshold for the entire 5 min, only then the alert will be raised. Or does it mean that Azure will monitor the metric in a 5 min window and if the metric goes above the threshold in that window, the alert will be raised.
For discrete metrics like Server Exceptions : For metrics like Server Exceptions which are discrete in nature, if I set the period to be 5 min, does it mean that the metric is summed and in any 5 min window if the sum reaches above the threshold then the alert will be raised.
Can anyone please elaborate on my above understanding and confirm if I am correct/wrong?