Percentile metrics provide a means to quantify the quality of service customers are receiving based on the response time that a fixed percentage (usually 95 percent) of users receive. These metrics are commonly referred to as 95th Percentile metrics. The exact percentage the system uses is configurable by the user.
To calculate this metric, the system uses a statistical method involving standard deviation rather than by tracking each individual event in memory (that approach would be prohibitive in terms of processing requirements and memory consumption). The percentile metric is calculated using this formula:
metric for information on how round-trip times are calculated.
This metric provides an estimate of the peak number of transactions per second that occur for a given service. This gives you the ability to evaluate the peak load that a service experiences and consider this when evaluating SLAs on the transaction itself. Some SLAs provide exemptions if concurrent activity exceeds a predetermined level. This metric provides a benchmark by which you can evaluate the peak levels of activity that a service experiences.
The algorithm for this metric records the termination time of each transaction as it completes. The termination time is used because that is the event that indicates the transaction was successfully completed. These times are then evaluated to determine the maximum number of transactions that complete in any given second.
metric for more information on service transaction counts based on arrival time.
metric for a description of how service processing times are calculated.