Despite the overwhelming popularity of Kubernetes as the default cloud native substrate, a significant majority of organizations continue to consume managed services on popular cloud providers to run critical applications. Often, the enterprise setting is a largely hybrid one with services spread across multiple hosting environments. With chaos engineering becoming an important part of DevOps strategy, there is a need to support fault-injection on several kinds of targets from a common control plane, while still maintaining homogeneity in experiment definition, & hypothesis validation. In this talk, the Litmus Maintainers will discuss how you can build chaos experiments for a hybrid environment wherein the chaos tooling is operated as a Kubernetes-native application whilst targeting remote cloud infrastructure and managed services. They will also discuss the various considerations in building out such a practice, especially around security and observability.