Service-direct: These faults run directly against an Azure resource, without any installation or instrumentation.You can organize faults to run in parallel or sequence, depending on your needs.Ĭhaos Studio supports two types of faults: A chaos experiment describes the faults to run and the resources to run against. Chaos experiments are the core of Chaos Studio. With Chaos Studio, you can orchestrate safe, controlled fault injection on your Azure resources. To check, you run chaos experiments as deployment gates in your continuous integration/continuous deployment pipelines. Then, you continuously validate that new deployments won't regress resilience. Validate that live site tooling, observability data, and on-call processes still work in unexpected conditions.įor many of these scenarios, you first build resilience by using ad-hoc chaos experiments.Build confidence in services built on cloud-native architectures.Ensure that services migrated from an on-premises or other cloud environment remain resilient to known failures.
Plan capacity needs for production environments.Develop application performance benchmarks.Run high-availability drills to test application resilience against region outages, network configuration errors, high-stress events, or noisy neighbor issues.Do business continuity and disaster recovery drills to ensure that your application can recover quickly and preserve critical data in a disaster.Prepare for a major event or season with "game day" load, scale, performance, and resilience validation.Ensure that post-incident repairs prevent the incident from recurring. Reproduce an incident that affected your application to better understand the failure.You can use Chaos Studio for the following common chaos engineering scenarios: You can do shift-left scenarios without any real customer traffic. Shift left: These scenarios can use a development or shared test environment.
Usually, you do shift-right scenarios with real customer traffic or simulated load.
0 kommentar(er)
