Krkn

Krkn is a free, self-hosted chaos-engineering toolkit for deliberately disrupting Kubernetes and OpenShift environments. It suits platform teams and chaos engineers who need cluster-level experiments tied to measurable reliability objectives. Its broad scenario catalogue and multiple execution methods make it adaptable, while the operational prerequisites and recovery risks demand disciplined use.

Overview

Krkn is designed to measure application or cluster resilience through controlled failures, not to provide security protection or compliance coverage. It is a Cloud Native Computing Foundation Sandbox project and supports Kubernetes, OpenShift, cloud fault injection, network fault injection, and blast-radius controls. The recommended operating model is to run it outside the cluster against the Kubernetes or OpenShift API.

Getting started requires Kubernetes or OpenShift 1.21 or newer, cluster-admin kubeconfig access, and Docker 20.10+ or Podman 4.0+. Those requirements make Krkn a better fit for teams that already control cluster administration than for developers seeking a lightweight local testing utility.

Key features

More than 20 scenarios cover pods, containers, nodes, networks, applications, resources, storage, and time. That range supports failure testing across infrastructure and workloads instead of constraining experiments to one class of outage.

krknctl discovers, runs, and orchestrates scenarios and is the preferred interface. Pre-built container images reduce packaging work, while a browser dashboard offers a graphical route and a standalone Python program provides another deployment path. The Python installation requires a Python 3.9 environment and manual dependency management, so it is the least convenient option for teams that want a tightly managed runtime.

Krkn can evaluate Prometheus PromQL queries against defined SLO pass/fail criteria and inspect metrics during and after a run. Dashboard integrations with Elasticsearch and Grafana, plus telemetry delivery to S3 or Elasticsearch, help teams retain and review experiment results.

Cerberus monitors Kubernetes and OpenShift health, exposes a go/no-go signal, and provides reports and a metrics API. Slack failure reporting is useful for operational workflows. Krkn-AI applies evolutionary algorithms to objectives such as SLOs and application health checks to seek impactful experiments, giving experienced teams a way to explore beyond manually selected scenarios.

Krkn saves recovery state before making changes and can automatically restore a cluster when chaos goes wrong. That safeguard is valuable, but the project also acknowledges that some disruptive scenarios can leave a target system non-recoverable; rollback should therefore be treated as protection, not a guarantee.

Pricing

The free plan has no stated seat, usage, or scenario quota on the opened plan information. There is no paid tier described, so organizations should plan to operate and maintain the software themselves rather than expect a hosted service or commercial support entitlement.

Platforms

Krkn is available through an API, Linux, self-hosted deployment, and a web dashboard. The CLI, containers, dashboard, and Python program let teams choose between automation, packaging convenience, browser management, and direct scripting.

Who it's for

Krkn is aimed at users ranging from first-time chaos engineers to teams building repeatable resilience-testing workflows. It is strongest for Kubernetes or OpenShift operators with cluster-admin access, Prometheus-based objectives, and the operational maturity to isolate blast radius and handle failed recovery. It is a poor match for teams that need a security product, formal compliance adherence, or a managed experiment service.

Pros and cons

Pros

• Broad failure coverage: More than 20 scenarios span compute, networking, applications, storage, resources, and time.
• Objective-based evaluation: PromQL-backed SLO pass/fail checks turn experiments into measurable reliability tests.
• Flexible operation: krknctl, containers, a dashboard, and Python support different deployment and automation preferences.
• Operational observability: Elasticsearch, Grafana, S3, Cerberus, metrics APIs, and Slack reporting support review and incident workflows.
• No software charge: The listed plan costs nothing.

Cons

• High privileges and setup: Kubernetes or OpenShift 1.21+, cluster-admin access, and container tooling raise the entry bar.
• Recovery is not assured: Some scenarios may leave systems non-recoverable despite saved state and automatic restoration.
• Execution constraint: krkn-hub runs one scenario type per execution; combining types requires standalone Krkn.
• Manual Python maintenance: The standalone route requires Python 3.9 and dependency management.
• No security or compliance role: Krkn says it is not a security provider and does not currently adhere to compliance standards.

Alternatives

Choose Reliably instead if you want an open-source approach that emphasizes avoiding lock-in across Linux, macOS, Windows, self-hosted, and web environments.

Chaos Toolkit is a free-forever, Apache-licensed open-source CLI for teams wanting a simpler general chaos-engineering tool.

ChaosBlade suits teams seeking an open-source tool with support for three system platforms, four programming-language application types, and more than 200 experiment scenarios.

Toxiproxy is the narrower choice for injecting TCP-network conditions through an open-source MIT-licensed proxy.

Chaos Mesh fits teams wanting an open-source, cloud-native platform that runs on Kubernetes.

Krkn-AI is the focused alternative for its open-source AI-driven chaos experimentation.

LitmusChaos offers either a free hosted SaaS option or an open-source platform installed on your own cluster.

Pumba is a free, Apache-licensed container chaos-testing tool for Linux and macOS.

Browse the Chaos Engineering Platforms list for more options.

Verdict

Choose Krkn if you operate Kubernetes or OpenShift and want a no-cost, self-hosted way to run diverse disruption experiments, connect them to Prometheus SLOs, and monitor outcomes through operational tooling. Look elsewhere if you need guaranteed reversibility, lower-privilege setup, built-in compliance coverage, or a managed service.