Fixing maximum heap size as a threshold test - quick-perf/doc GitHub Wiki

The principle of a threshold test is to compare a value in the current build to a threshold. If this value exceeds the threshold, the test will fail and we are alerted of a possible performance degradation.

In this article, Martin Fowler gives the example of the amount of time taken by a test.

When you set maximum heap size for a test with the help of @HeapSize or @Xmx, the test may fail one day because of an OutOfMemoryError. You may consider this as a performance issue or if not increase maximum heap size. So, fixing maximum heap size could be seen as a threshold test. A significant increase of heap allocation may also lead to a significant increase of the test time length because of garbage collection activity. So your build duration could increase, possibly alerting you of a significant increase of heap allocation for a test. @ExpectNoJvmIssue can help you to check that most of the test time is spent to garbage collect objects.