A common type of performance test engineers do is capacity testing, which aims at determining the maximum number of users or throughput the application can sustain before performance requirements (possibly corresponding to Service Level Objectives, or SLOs) are no longer met. In such situations, performance engineers spend a lot of time analyzing load test results, for example, to identify the maximum throughput while some SLOs are met like maximum response time or error rates.
Akamas already featured automated performance scoring (known as windowing), for example, to automatically discard performance tests warm-up and tear-down periods and compute an accurate experiment score.
In this release, Akamas automated scoring has been enhanced to better support performance constraints. For example, you can create a study with an optimization goal set to maximize system throughput while the response time is below 100 ms and the error rate < 1%. Akamas analyzes performance test metrics (e.g. from JMeter, NeoLoad, or LoadRunner) and automatically identifies the performance score as the maximum throughput the system reached while your SLOs are not violated (see the following picture from the Akamas UI).