Chaos Toolkit

Chaos Toolkit is a free, open-source toolkit for developers who want to explore systems and look for weaknesses through chaos engineering. An experiment defines a steady-state hypothesis and probes to check whether the system meets it, along with a method made of actions and probes. If the steady state is not met, the method is not applied and the experiment must stop. Experiments can be saved as JSON or YAML, making them suitable for collaboration and orchestration alongside code. The toolkit extends to systems through its Open API and can be embedded in a CI/CD chain. Listed integrations include AWS, Azure, Google Cloud Platform, Kubernetes, Kafka, Slack, Datadog, Grafana, OpenTelemetry, and Prometheus. The standalone CLI is licensed under Apache License 2.0 and is implemented in Python 3, officially supporting Python 3.8 and later; it has only been tested against CPython. Installation guidance covers macOS, Debian or Ubuntu, and Windows. Users can add extensions after installing the CLI and core library. Reliably is a separate platform, not the standalone CLI.

Chaos Toolkit gives developers a code-oriented way to define experiments, with a steady-state check that can stop a method from running. It is a free standalone CLI; Reliably is a separate offering.

It suits developers who want to define and run system experiments as code, including in a CI/CD chain. Users should be comfortable with Python 3 and installing extensions for different areas of chaos engineering.

Chaos Toolkit is an open-source tool for developers who want to probe system resilience through experiments they can keep with their code. It is best suited to teams comfortable installing and operating a Python command-line tool; teams wanting a managed chaos engineering service should look elsewhere.

Overview

Experiments are written in JSON or YAML, making them practical to collaborate on and orchestrate alongside other code. Each defines a steady-state hypothesis, probes to check it, and a method of actions and probes. If the system fails the steady-state check, the method does not run and the experiment must stop. That is a useful precondition for fault testing, but it is a guardrail rather than a substitute for careful experiment design.

The toolkit is an Apache License 2.0 project with a standalone, self-hosted CLI. The project welcomes contributions through issue reports, voting, stars, Slack, pull requests and extension proposals; its contributor guide recommends enabling two-factor authentication on contributor accounts.

Key features

Experiments, automation and extensions

Chaos Toolkit can be embedded in a CI/CD chain, so teams can incorporate experiments into their delivery workflow. Its Open API supports extensions for different systems, installed after the CLI and core library. Integrations include AWS, Azure, Google Cloud Platform, Kubernetes, Kafka, Slack, Datadog, Grafana, OpenTelemetry and Prometheus.

It supports Kubernetes, cloud and network fault injection, experiment scheduling and blast-radius controls. That breadth is useful for teams building a varied experiment practice, though the extension-based model means users must install and manage the pieces they need.

Python CLI

The CLI is built in Python 3, officially supports Python 3.8 and later, and has been tested only with CPython. Installation uses pip, with setup instructions for macOS, Debian or Ubuntu, and Windows. Teams already comfortable with Python can adopt a conventional command-line workflow; those seeking a graphical, managed experience may prefer another product.

Pricing

Chaos Toolkit: free. The plan is free forever and includes the open-source CLI under Apache License 2.0. There is no paid tier to weigh against it in the standalone offering. Reliably is a separate platform, not a paid Chaos Toolkit plan, and has SaaS, private SaaS and on-premises options.

Platforms

Chaos Toolkit is listed for API, Linux, macOS, self-hosted, web and Windows. Its CLI setup instructions cover macOS, Debian or Ubuntu, and Windows; the runtime itself requires Python 3.8 or later and has only been tested against CPython.

Who it's for

This is a strong fit for developers and platform teams that want versionable experiments, a steady-state check before actions run, and the flexibility to add integrations or wire experiments into CI/CD. It is less suitable for teams that want a hosted, turnkey service or would rather not operate a Python CLI and extensions themselves.

Pros and cons

• Pro: JSON or YAML experiment definitions can live with other code, supporting familiar collaboration and orchestration.
• Pro: A failed steady-state check prevents the method from running, giving experiments a clear safety gate.
• Pro: The Open API, extension directory and CI/CD support allow teams to adapt experiments to their systems and workflow.
• Con: The standalone tool is self-hosted and Python-based, so teams must handle installation and operation rather than rely on a managed Chaos Toolkit service.
• Con: Extensions are installed separately after the CLI and core library, adding setup work for teams needing multiple facets of chaos engineering.

Alternatives

Chaos Engineering Platforms is a useful starting point for comparing options across the category.

• Chaos Mesh is another free, self-hosted option; choose it if a cloud-native chaos engineering platform that runs on Kubernetes is the better fit.
• Reliably is worth considering if you want a separate platform with SaaS, private SaaS or on-premises deployment options. Its Free plan is free for 1 team member.
• ChaosBlade is another free, open-source chaos engineering tool, with support for 3 system platforms, 4 programming-language applications and 200+ experiment scenarios.
• Gremlin may suit teams considering a paid product with a trial: its Platform trial is free, billed as a 30-day free trial, requires no credit card and does not auto-convert.
• Steadybit is a paid alternative with a Professional plan and an Enterprise plan that adds on-prem installation, SAML, audit log and enterprise support.
• PowerfulSeal is a free, self-hosted Kubernetes chaos testing tool installable with pip or Docker.
• LitmusChaos offers both free hosted SaaS and an open-source platform for installation on your own cluster.
• AWS Fault Injection Service is a paid alternative billed at $0.10 per action-minute, with additional account and GovCloud rates.

Verdict

Choose Chaos Toolkit if your team wants free, code-oriented experiments, a pre-run steady-state safeguard and extensibility in a self-managed Python CLI. Look elsewhere if operating that toolchain is a burden or you need chaos engineering delivered as a managed service.