Abstract: A method and apparatus may rapidly test applications by causing or simulating failures within nodes of a cloud computing system in support of both application and infrastructure testing. The method and system may support a variety of “attacks” including the ability to stop or freeze application servers, insert latency or drop packets between servers, constrain CPU or memory, and disable various software flows and applications. Rather than randomly inserting random failures or simulated failures into cloud-based computing system nodes to test their durability and the efficacy of particular applications or services that are executing within the system, the system and methods include a user interface for manually controlling the system attacks.