Krkn vs Krkn-AI

Krkn injects deliberate failures into Kubernetes clusters so teams can examine resilience under disruption. Its 20-plus scenarios cover pods, containers, nodes, networks, applications, resources, storage and time. The recommended krknctl command-line tool discovers, runs and coordinates scenarios; other choices include pre-built containers, a browser dashboard and a standalone Python program. Krkn evaluates Prometheus metrics against SLO pass/fail criteria and can review metrics during and after scenarios. Dashboard integrations include Elasticsearch and Grafana, and telemetry can be sent to S3 or Elasticsearch. Cerberus monitors Kubernetes and OpenShift health, providing a go/no-go signal, reports and a metrics API; it can also report failures in Slack. Krkn-AI uses evolutionary algorithms against user-defined objectives to identify experiments. Krkn says it saves recovery state and can restore a cluster, but some disruptive scenarios may leave the target unrecoverable. It is intended for resilience measurement, not security provision, and does not currently adhere to compliance standards. Requirements include Kubernetes or OpenShift 1.21+, cluster-admin kubeconfig access, and Docker 20.10+ or Podman 4.0+.

Krkn-AI is a free, open-source tool for automatically running chaos scenarios and finding experiments to evaluate system resilience. It uses evolutionary algorithms to generate experiments from Krkn scenarios, guided by user-defined objectives such as SLOs and application health checks. Its discover command creates a YAML configuration populated with cluster component information and basic setup. Results show the impact of scenarios on application liveness and include fitness scores; Prometheus metrics can measure system response. An interactive dashboard can monitor experiment results. The project describes testing Kubernetes and OpenShift applications and requires a cluster kubeconfig. Krkn-AI can be installed as a Python CLI or run from a pre-built container image, including on Kubernetes. Listed requirements include Python 3.11 or later, krknctl, oc or kubectl, Podman, Helm 3.x, and a Kubernetes or OpenShift cluster kubeconfig. The repository lists the Apache-2.0 license. Its README describes the tool as under active development and advises users to use it at their own risk. The standard container run example uses host networking and privileged mode, while the guide also documents a FUSE-based option without the privileged flag.