Once you have a metric clearly defined, you need to define the criteria based on that metric that will make the benchmark validated or rejected—is your application fast enough to rephrase it at a high level?
Generally, it is a number that can be expressed as a time unit or percentage, depending on the metric. If the measure is lower (or higher) than this number then the benchmark is rejected.
Most of the time, the metric is not self-sufficient and needs some additional parameters to be able to define the acceptance criteria in a measurable way. Here are some common examples:
- The client execution duration must be under 1 second for 64 concurrent users
- The client latency must be under 250 milliseconds when 128 ...
