The average is determined by the sum of time noted for a procedure name to run divided by the total number of times the procedure ran. The data used to determine this average is collected during a polling interval. This calculated result is then compared to the threshold defined for the procedure name.

Since this calculation determines an average based on data collected for each procedure, an individual spike will not necessarily skew the average. The rule's algorithm is designed to eliminate these spike conditions, minimizing unnecessary alerts.

Note: The Procedure Duration High rule measures the execution time of the ABL procedure only from the server's, or agent's, viewpoint. The time measure does not include network and client processing overhead.