krkn-operator

Krkn Operator is a Kubernetes-native platform built on the Krkn framework that provides centralized, multi-cluster chaos engineering orchestration for Kubernetes and OpenShift environments. It allows teams to run chaos experiments across multiple clusters from a single control plane, making it easier to validate system resilience at scale. Key capabilities include multi-cluster orchestration, where chaos experiments can be executed across both Kubernetes and OpenShift clusters simultaneously. The platform features a Chaos Studio that lets users visually compose and execute reusable chaos workflows without writing code. Access control is built in, allowing administrators to manage users, groups, cluster access, and permissions. Integration with Open Cluster Management (OCM) and Red Hat Advanced Cluster Management (ACM) enables discovery and execution of experiments on clusters managed by these platforms. Installation is straightforward via Helm, with a single command deploying the operator into a dedicated namespace. The project also publishes OLM/OperatorHub bundle images for both generic Kubernetes and OpenShift, with separate channels (stable-kubernetes and stable-ocp). The OLM bundle declares Kubernetes 1.19.0 as its minimum version. On OpenShift, the bundle intentionally does not create Route, Ingress, or Gateway resources, leaving TLS and networking configuration to the cluster administrator. The ecosystem includes krkn-operator-console, a web console and Chaos Studio for the operator, and krkn-operator-acm, which provides OCM and Red Hat ACM integration for multi-cluster environments. The API supports scenario-based experiment execution where requests identify a scenario by name rather than supplying an executable image directly. The operator resolves the image from the scenario name and selected registry. Private registries are supported by setting the private flag and including the registry name. Image signature verification can be toggled by administrators via the API, with verification remaining observable even when enforcement is disabled. The project is licensed under Apache License 2.0 and welcomes contributions through its CONTRIBUTING.md guidelines.