The Target SystemThe Platform: A Three-Node Kubernetes ClusterThe Application: A Single Service, Replicated Three TimesThe People: Application Team and Cluster AdministratorsHunting for a WeaknessNaming Your ExperimentDefining Your Steady-State HypothesisInjecting Turbulent Conditions in an Experiment’s MethodUsing the Kubernetes Driver from Your MethodBeing a Good Citizen with RollbacksBringing It All Together and Running Your ExperimentOvercoming a Weakness: Applying a Disruption BudgetSummary