Performance tests are executed to establish how quickly an application can perform an action, given a set of resources. Performance tests are often executed using specialized tools and run against a fully assembled system. If the tools used for automated API or UI tests record the duration of a test, the duration of these tests can be used as performance results as well.

To compare results over multiple test runs, it is important to ensure that all factors influencing performance are kept the same between tests. The setup of virtual machines for both test subjects and test runners should stay the same. The application configuration should remain constant and integration points should be in the same state as much as possible—for example, instead of reusing the same database, the same database should be restored from a backup before every performance test. This ensures that the results are comparable.

While performance and load tests are often mixed up, they are two different types of tests.