Skip to content

Adoption & Ecosystem

Who uses it

The repository has no ADOPTERS file; the README asks companies to register adoption through a GitHub issue instead. The only named production adopters with a citable source come from Alibaba's own platform write-up, which states that more than 40 companies had registered and that the following used it in production.

OrganisationUse caseSource
ICBC (Industrial and Commercial Bank of China)Production chaos engineeringAlibaba Cloud blog
China MobileProduction chaos engineeringAlibaba Cloud blog
XiaomiProduction chaos engineeringAlibaba Cloud blog
JD.comProduction chaos engineeringAlibaba Cloud blog

Beyond these four named companies, no further citable adopters were found. Treat the "40+ companies" figure as the vendor's own count.

Adoption signals

Measured on 2026-06-27 from the GitHub API (gh api repos/chaosblade-io/chaosblade): 6,358 stars, 1,001 forks, 348 open issues, and roughly 52 contributors. CNCF lists the project at Sandbox maturity, accepted on 2021-04-28. LFX Insights reports a concentration risk: in a recent quarter around 15 contributors were active and one contributor accounted for more than half of the activity, a single-maintainer dependency. The 2025 survey paper (arXiv 2505.13654) describes ChaosBlade as a Sandbox project under continued development across its release history.

Ecosystem

Chaosblade is the CLI at the centre of a family of repositories:

  • chaosblade-exec-os, chaosblade-exec-jvm, chaosblade-exec-cplus, and the other chaosblade-exec-* repos: per-domain executor binaries and their YAML specs, which the CLI build clones and packages.
  • chaosblade-spec-go: the experiment-model SDK that defines spec.ExpModel, spec.Executor, and the YAML parsing used by the CLI.
  • chaosblade-operator: the Kubernetes operator that exposes chaos experiments as a Custom Resource Definition (CRD).
  • chaosblade-box: the platform UI, deployed via Helm, with Prometheus integration and the ability to host LitmusChaos experiments as well.
  • blade-ai: a newer Python agent layer that drives experiments from a plain-language failure description.

Alternatives

Chaosblade's distinguishing trait is breadth: one CLI and one experiment model span host, JVM, C++, Docker, CRI, Kubernetes, and cloud, with strong application-layer injection (JVM via a Java agent, C++ via GDB). The Kubernetes-native alternatives concentrate on cluster-level chaos driven by custom resources.

AlternativeDiffers by
Chaos Mesh (CNCF Incubating)Kubernetes-native, CRD-driven; originated at PingCAP; centred on cluster chaos rather than multi-platform CLI
LitmusChaos (CNCF Incubating)Kubernetes chaos workflows and an experiment hub; declarative pipelines over a single CLI
GremlinCommercial SaaS with a hosted control plane; not open source
AWS Fault Injection Service (FIS)Managed fault injection scoped to AWS resources
Chaos ToolkitOpen, extensible JSON/YAML experiment format with a plugin ecosystem; provider-agnostic but lighter on built-in injectors