I am trying to set up an incident rule for detection of a specific exception that is thrown when connection to an application is lost. I have created a measure for the metric "Count exception" and specified the name of the exception class and which message the exception should contain. My question is as follows: Which aggregation tupe should be chosen if an incident should be raised when the exception has been thrown a least 10 times in the last 5 minutes? The evaluation timeframe is set to 5 minutes and the severe threshold to 10 but I don't know if the aggregation type should be 'sum' or 'count'. What is the diffference?
See this post from earlier today:
Basically, sum will be the total sum of the exceptions occurrences, whereas count will be the count of PurePaths that went into calculating the measure. So in this case you will also want to use sum.