CheckOps home

Tools directory

L

CheckOps

LitmusChaos

LitmusChaos is a 100% open-source CNCF chaos engineering platform that uses Kubernetes custom resources to define chaos and steady-state hypotheses, with a hub of experiments and a control-plane UI.

Visit website

Latest public signals

Not refreshed yet. Refresh reads the public site and, when the entry has a GitHub repository, the latest release.

Visit Website

Introduction

LitmusChaos is an open-source chaos engineering platform that helps teams find weaknesses and potential outages in infrastructure by running chaos tests in a controlled way. The README describes it as 100% open source and a CNCF project. It is Apache-2.0 with about 5,700 GitHub stars.

Litmus takes a cloud-native approach: the platform runs as microservices and uses Kubernetes custom resources to define both the chaos intent and the steady-state hypothesis, so experiments check whether the system stays healthy while faults are injected.

A control plane (ChaosCenter) provides a UI for creating, scheduling and monitoring experiments, and ChaosHub offers reusable experiments that the community shares.

In the quality loop, Litmus lets SREs and developers run resilience tests in staging and production, often as part of game days or scheduled reliability checks.

Features

  1. Chaos experiments as CRs: Kubernetes-native definitions.
  2. Steady-state hypothesis: Probes check health during experiments.
  3. ChaosCenter: Web control plane for experiments and results.
  4. ChaosHub: Reusable, community-shared experiments.
  5. Scheduling: Run experiments on a schedule.
  6. Community translations: README and docs in several languages.
  7. Environments: Connect the control plane to the clusters you test; see the docs.
  8. CI/CD use: Trigger experiments from pipelines.

Pros

  1. Fully open source: CNCF governance.
  2. Hypothesis-driven: Built-in probes.
  3. Reusable experiments: Hub reduces setup.
  4. Free: Apache-2.0.
  5. Active: Commits in November 2026.

Cons

  1. Kubernetes-centric: Best for Kubernetes.
  2. Platform to run: Control plane needs hosting.
  3. Learning curve: Several concepts to learn.
  4. Blast radius: Needs careful scoping.
  5. Depends on observability: Probes need good signals.

Install the ChaosCenter using the docs (Helm or manifests), connect a cluster, then pick an experiment from ChaosHub, such as pod delete, and add a probe that checks your service health endpoint during the run.

Tips: define the steady state first (for example error rate under 1%), start with a single replica in staging, and record outcomes so you can track resilience over time.

Resources:

Pricing

LitmusChaos is free and open source under Apache-2.0.

  1. SRE teams running game days on Kubernetes.
  2. Developers adding resilience tests to CI.
  3. Organisations wanting a shared experiment hub.
  4. Not a fit for non-Kubernetes estates.

Conclusion

LitmusChaos offers hypothesis-driven chaos engineering with a control plane and reusable experiments.

Define steady states and scope experiments carefully as you expand.

Alternatives

Other tools on CheckOps that cover similar ground. Open a page, or compare one with LitmusChaos side by side.