This presentation discusses how test methods should be extended to cope with evolving systems. It differentiates between evolution at design time and at runtime and the various test kinds used to assure system quality after system changes. Runtime evolution in particular requires so called online tests, which support the active evaluation of systems in their production environments as pre-deployment tests in experimental environments are limited. Runtime models and built-in tests are a basis for online tests and allow to check for system sanity, performance degradation, and alike.