Azure Chaos Studio

Azure Chaos Studio is a fully managed chaos engineering experimentation platform. It improves application resilience by deliberately introducing faults that simulate real-world outages, from late-stage development through production.

Experiments combine service-direct faults applied to Azure resources, such as shutting down virtual machines or failing over databases, and agent-based faults that run inside virtual machines to consume CPU or memory, or disrupt the network. Faults are organised in steps and branches, targets are selected explicitly, and access is controlled through Azure permissions. Experiments can be automated in CI/CD pipelines.

Azure Chaos Studio is a Microsoft Azure service billed according to usage.