Adoption & Ecosystem
Who uses it
The repository's ADOPTERS.md lists more than 40 organizations. The entries below are the ones with a public, citable account of how they use Chaos Mesh. ADOPTERS entries are self-reported, so the depth of "production use" varies by how each company described it.
| Organisation | Use case | Source |
|---|---|---|
| ByteDance | Integrated Chaos Mesh as the fault injection engine under its internal chaos platform. | ADOPTERS.md |
| Tencent (Interactive Entertainment) | Fault isolation and service degradation testing after migrating to TKE. | Chaos Mesh blog |
| DataStax | Built Chaos Mesh into Fallout for testing AstraDB, its Cassandra-based DBaaS. | Talk |
| Authzed | Uses TimeChaos to fake vDSO clock time when testing SpiceDB. | Talk |
| Percona | Chaos testing of Percona Kubernetes Operators. | Percona blog |
| NetEase Fuxi Lab | Improving the stability of its internal hybrid cloud. | ADOPTERS.md |
The CNCF project page also lists ByteDance, NetEase, PingCAP, Tencent, and Microsoft Azure among adopters. Other names in ADOPTERS.md include RabbitMQ, GreptimeDB, KingNet, DigitalChina, Xpeng, Maycur, Prudential, and Qiniu.
Adoption signals
Measured on 2026-06-24 from the GitHub REST API for repos/chaos-mesh/chaos-mesh:
- Stars: 7,763
- Forks: 1,007
- Open issues: 549
- Contributors: roughly 221 (including anonymous)
- Created: 2019-09-04; last push: 2026-06-23
A comparative study (arXiv 2505.13654) counts 74 Chaos Mesh releases against 106 for LitmusChaos and frames the two as the leading Kubernetes-native chaos tools with sustained long-term adoption.
Ecosystem
- Microsoft Azure Chaos Studio integrates Chaos Mesh faults to run experiments against AKS, including pod kill (Azure docs).
- Distribution through the KubeSphere App Store and the Civo Kubernetes marketplace.
- Built-in Prometheus metrics, plus Workflow (Argo-style experiment orchestration) and Schedule (cron) bundled in the project itself.
Alternatives
| Alternative | Differs by |
|---|---|
| LitmusChaos (CNCF) | ChaosHub ships 50+ prebuilt experiments and broad multi-cloud coverage, but uses a two-stage ChaosEngine + ChaosExperiment CRD model that is more verbose. |
| Chaos Toolkit | Generic, declarative chaos framework that is not Kubernetes-specific. |
| Gremlin / Steadybit | Commercial SaaS chaos platforms. |
| AWS FIS | AWS-managed fault injection, scoped to AWS resources. |
| Chaos Monkey / Simian Army | Netflix's platform-agnostic originals, focused on instance termination rather than fine-grained faults. |
Pick Chaos Mesh when you are on Kubernetes, want one CRD per fault type for a clean kubectl apply and GitOps fit, and need its low-level faults (TimeChaos, JVMChaos, KernelChaos, IOChaos) for testing distributed consensus, locks, or TTL caches. The trade-off is that it assumes a privileged DaemonSet and is therefore invasive to the cluster, with higher operational cost (Container Solutions comparison).
Sources
- ADOPTERS.md: https://github.com/chaos-mesh/chaos-mesh/blob/master/ADOPTERS.md
- Chaos Mesh project page (CNCF): https://www.cncf.io/projects/chaosmesh/
- GitHub REST API repos/chaos-mesh/chaos-mesh: https://api.github.com/repos/chaos-mesh/chaos-mesh
- Microsoft Azure Chaos Studio + Chaos Mesh (AKS): https://docs.microsoft.com/en-us/azure/chaos-studio/chaos-studio-tutorial-aks
- LitmusChaos vs Chaos Mesh (reintech): https://reintech.io/blog/litmuschaos-vs-chaos-mesh-kubernetes-chaos-tool-comparison-2026
- Comparing Chaos Engineering Tools (Container Solutions): https://blog.container-solutions.com/comparing-chaos-engineering-tools
- Chaos Engineering in the Wild (arXiv 2505.13654): https://arxiv.org/html/2505.13654v1