Cloud architectures are incredibly complex and, often, it's all but impossible to predict and prevent failure scenarios. Enter the ascendant discipline of chaos engineering, which aims to discover cloud failure points, in in-production systems, before they become disasters.