Suggest a tool

Category guides · published

"Best open source chaos engineering tools": 10 listed, sorted by GitHub stars

Metrics as of , from the GitHub or GitLab API of each repository. Refreshed monthly.

The chaos category covers fault injection and resilience testing tools, following the definition in the FreeQATools category list. The listed tools differ in where they inject faults: Kubernetes clusters, containers, hosts and applications, or the network path between services. The table below is sorted by GitHub stars on the fetch date.

Target environments

Chaos Mesh defines experiments as Kubernetes custom resources (README). LitmusChaos combines custom resources such as ChaosExperiment and ChaosEngine into workflows (README). Krkn runs its scenarios against Kubernetes and OpenShift clusters (docs). kube-monkey deletes pods at random in a Kubernetes cluster (README).

Pumba works on Docker, containerd and Podman containers on Linux (README). ChaosBlade documents both host environments and Kubernetes clusters (docs). Its README also lists method-level injection into Java and C++ applications (README).

Chaos Monkey terminates instances and containers on Spinnaker backends such as AWS, Google Compute Engine and Kubernetes (README). Apps must be managed with Spinnaker for Chaos Monkey to act on them (Chaos Monkey). Chaos Toolkit positions itself for cloud environments, datacenters and CI/CD (README). It reaches target platforms through extensions (Chaos Toolkit).

Faults at the network and API layer

Toxiproxy acts on TCP connections that an application routes through the proxy (README). Its documented toxics include latency, down, bandwidth, timeout and packet_loss (README). Tests change the state of those connections over an HTTP API (Toxiproxy).

Hoverfly is listed in both the chaos and test data categories (Hoverfly). It simulates the APIs an application depends on (Hoverfly). Its simulations can inject network latency, random failures and rate limits (README).

Experiment formats and CI

Chaos Mesh writes experiments as Kubernetes custom resources (README). LitmusChaos also writes experiments as Kubernetes custom resources (README). Chaos Toolkit uses JSON experiment files (docs). Pumba takes CLI commands with flags, such as kill, netem and stress (README). kube-monkey reads opt-in labels on Kubernetes app manifests (README).

Chaos Monkey takes per-application settings in the Spinnaker web UI (docs). ChaosBlade accepts CLI commands, an HTTP server mode and YAML CRDs on Kubernetes (docs).

Chaos Mesh documents a GitHub Actions integration, chaos-mesh-action (docs). LitmusChaos documents the GitHub Action litmuschaos/github-chaos-actions (README). Chaos Toolkit documents a GitHub Action run from a workflow (docs). Krkn documents krkn-hub container images for CI/CD systems such as Jenkins and GitHub Actions (docs).

For ChaosBlade, Hoverfly, kube-monkey and Pumba, the CI row of the comparison table reads not documented, because the fetched docs did not describe a CI integration (methodology).

How this list is sorted

The quoted title is a search query, not a verdict. FreeQATools does not rank tools by opinion. The list below is sorted by GitHub stars, descending, as fetched from the repository host API on the date shown with each value.

Stars count how many accounts have starred a repository. They say nothing about fit for a given project, so the documented facts under each tool are the part to compare.

Every tool here meets the inclusion rules: an OSI-approved license, a public repository, software testing or quality as its primary purpose and at least one release or tag. Each status badge follows the status rules on the methodology page.

Open-source chaos and resilience tools by GitHub stars

Chaos and resilience tools. Sorted by GitHub stars, descending. Select a column heading to change the sort.
Chaos Monkey17,14512025-01-061v2.1.32024-10-031Apache-2.0Goinactive
Toxiproxy12,35612025-03-181v2.12.02026-08-251MITGoactive
Chaos Mesh7,91112026-08-181v2.8.42026-09-061Apache-2.0Goactive
ChaosBlade6,52112026-09-211blade-ai-v0.7.22026-07-281Apache-2.0Pythonactive
LitmusChaos5,61912026-09-1713.32.02026-09-221Apache-2.0Goactive
Pumba3,17112026-08-2311.2.12026-08-271Apache-2.0Goactive
kube-monkey3,08012026-09-201v0.7.02026-09-201Apache-2.0Goactive
Hoverfly2,52212026-09-211v1.12.152026-09-211Apache-2.0Goactive
Chaos Toolkit2,02812026-08-0811.20.02026-08-091Apache-2.0Pythonactive
Krkn50312026-09-031v5.2.92026-09-221Apache-2.0Pythonactive

1 Fetched from the GitHub or GitLab API on . Hover a value for its own date.

Filters for language, license and status are on the chaos and resilience category page.

Tools in this list

Each entry gives the tool's one-line summary from its README and the facts its documentation states for this category, each with its source. Facts that are not documented are left out here and marked on the comparison pages.

  1. Chaos Monkey

    Netflix chaos tool that randomly terminates instances and containers in production, integrated with Spinnaker. README, read 2026-09-22

    Target environments
    Spinnaker backends: AWS, Google Compute Engine, Azure, Kubernetes, Cloud Foundry source: README
    Fault types
    Random termination of virtual machine instances and containers source: README
    Experiment format
    Per-application settings in the Spinnaker web UI source: Docs: Configuring behavior via Spinnaker
    CI integration
    Integrated with Spinnaker, a continuous delivery platform source: README
    Install method
    Go (go get github.com/netflix/chaosmonkey/cmd/chaosmonkey) source: README

  2. Toxiproxy

    TCP proxy with an HTTP API for simulating latency, outages and other network faults in tests and CI. README, read 2026-09-22

    Target environments
    TCP connections between an application and its services, routed through the proxy source: README
    Fault types
    latency, down, bandwidth, slow_close, timeout, reset_peer, slicer, limit_data, packet_loss source: README
    Experiment format
    HTTP API, CLI and client libraries (Ruby, Go, Python, .NET, PHP, Node.js, Java and others) source: README
    CI integration
    Designed for testing, CI and development environments source: README
    Install method
    Docker image ghcr.io/shopify/toxiproxy source: Docs: GitHub container package toxiproxy

  3. Chaos Mesh

    Chaos engineering platform for Kubernetes that defines fault injection experiments as custom resources. README, read 2026-09-22

    Target environments
    Kubernetes; remote clusters managed from a management cluster source: README
    Fault types
    Pod, network, DNS, HTTP, I/O, time, stress, kernel, block device, JVM, physical machine, AWS, Azure, GCP source: README
    Experiment format
    Kubernetes custom resources; web dashboard and API source: README
    CI integration
    GitHub Actions (chaos-mesh-action) source: Docs: Integrate Chaos Mesh to GitHub Actions
    Install method
    Helm chart (helm repo add chaos-mesh https://charts.chaos-mesh.org) source: Docs: Install Chaos Mesh using Helm

  4. ChaosBlade

    Chaos engineering toolkit from Alibaba for injecting faults into hosts, containers, Kubernetes, Java and C++ applications. README, read 2026-09-22

    Target environments
    Host environments and Kubernetes clusters source: Docs: ChaosBlade introduction
    Fault types
    CPU, memory, network, disk, process; Java and C++ method-level injection; container and Pod kill source: README
    Experiment format
    CLI commands, HTTP server mode, YAML CRDs on Kubernetes, ChaosBlade-Box UI source: Docs: ChaosBlade introduction
    Install method
    Release toolkit download; chaosblade-operator Helm chart for Kubernetes source: README

  5. LitmusChaos

    Chaos engineering platform for Kubernetes that defines experiments as custom resources and runs them as workflows. README, read 2026-09-22

    Target environments
    Kubernetes resources; cloud platforms such as AWS, GCP and Azure; VMware source: Docs: Litmus Experiments
    Fault types
    Pod chaos: container kill, disk fill, pod delete, CPU, memory and IO stress, DNS errors, network latency, loss and corruption source: Docs: Litmus Experiments
    Experiment format
    Kubernetes custom resources (ChaosExperiment, ChaosEngine) combined into workflows source: README
    CI integration
    GitHub Action litmuschaos/github-chaos-actions source: Docs: GitHub Action for Chaos Engineering in Kubernetes
    Install method
    Helm 3 chart (litmuschaos/litmus) or kubectl YAML spec file source: Docs: ChaosCenter installation

  6. Pumba

    Chaos testing CLI that kills containers, injects network faults and stresses resources on Docker, containerd and Podman. README, read 2026-09-22

    Target environments
    Docker, containerd and Podman containers on Linux source: README
    Fault types
    Container kill, stop, pause and remove; network delay and packet loss; CPU, memory and IO stress source: README
    Experiment format
    CLI commands with flags (kill, netem, iptables, stress) source: README
    Install method
    Release binary or Docker image ghcr.io/alexei-led/pumba source: README

  7. kube-monkey

    Chaos Monkey implementation for Kubernetes that randomly deletes pods of opted-in applications on a configured schedule. README, read 2026-09-22

    Target environments
    Kubernetes clusters source: README
    Fault types
    Random pod deletion source: README
    Experiment format
    Opt-in labels on Kubernetes app manifests (kube-monkey/enabled, kube-monkey/mtbf) source: README
    Install method
    Helm chart kubemonkey/kube-monkey source: README

  8. Hoverfly

    API simulation tool that stands in for service dependencies, with latency and failure injection, a CLI and REST API. README, read 2026-09-22

    Target environments
    Linux, macOS and Windows binaries; Docker; Kubernetes via Helm source: Docs: Download and installation
    Fault types
    Network latency, random failures, rate limits source: README
    Experiment format
    Simulation JSON files (captured traffic, exported, edited and imported) source: Docs: Simulations
    Install method
    Binary archives; Homebrew (brew install SpectoLabs/tap/hoverfly); Docker image spectolabs/hoverfly; Helm source: Docs: Download and installation

  9. Chaos Toolkit

    Python command-line tool for writing and running chaos engineering experiments, extended to target platforms through drivers. README, read 2026-09-22

    Target environments
    Cloud environments, datacenters and CI/CD source: README
    Experiment format
    JSON experiment files source: Docs: Experiment
    CI integration
    GitHub Action run from a GitHub Workflow source: Docs: GitHub Action
    Install method
    uv or pip (uv tool install chaostoolkit) source: README

  10. Krkn

    Chaos and resiliency testing tool that injects pod, node, network and other failures into Kubernetes clusters. README, read 2026-09-22

    Target environments
    Kubernetes and OpenShift clusters; node scenarios through cloud APIs source: Docs: Chaos Scenarios
    Fault types
    Pod, container and node failures; node CPU, memory and IO hogs; network latency, packet loss and bandwidth limits source: Docs: Chaos Scenarios
    Experiment format
    Pre-built scenarios run with the krknctl CLI, or krkn-hub container images configured by environment variables source: Docs: Installation
    CI integration
    krkn-hub container images for CI/CD systems such as Jenkins and GitHub Actions source: Docs: Installation
    Install method
    krknctl CLI; krkn-hub container images; standalone Python program from Git source: Docs: Installation

Comparisons in this category