With the goal of validating system upgrades, we can up with a design pattern and logic on how to test the effects of events on a system. This requires a new way of adressing testing, but also how events could be modelled and injected into a scenario. This model applies to cases such as: * System Upgrades * Cloud migrations * Sytem shutdown * Network failures Some of these cases are new to many, but some are also known as Chaos Testing and Monkey Testing. Many of the methods we present here are already in use at Adobe, but in the case of chaos testing it is new. The model we present here, although working well with standard testing frameworks, requires a new way of looking at tests and how they are designed. We introduce three paradigms: * A test scenario should be interruptible * A test scenario when testing an event needs to simulate it at all of its stages of test execution.