High Velocity Development of Microservices
summary
Claims a way to move “testing in production” into CI by capturing real API traffic with trace IDs, replaying it against new service versions, and simulating stateful backends so only services under test run. Uses tracing tools like Jaeger/Zipkin/APM, a golden capture of live requests, and trace-id driven simulation to return exact historical responses; demo uses Mesh Dynamics to produce diffs, file Jira bugs, update goldens, and achieves 100% functional regression coverage with far less test scripting.
AI-written summary · about
chapters (9)
- 0:00Intro
- 0:39Current trends
- 1:11Typical testing goals
- 2:00Current testing approaches
- 3:14Newer approaches
- 6:17Reconciling state mismatch: options
- 9:57Testing in near-production
- 11:53Demo
- 19:43Outro
related talks
31mHundreds of Microservices without breaking your APIs
31mUnderstanding Microservices with Distributed Tracing
27mEffectively Testing Services
39mObservable microservices
52mPerfecting reliable code delivery for the cloud
57mEnd-to-End Automated Testing in a Microservices Architecture
35mMicroservices, Service Mesh, and CI/CD Pipelines: Making It All Work Together [I]
46mObserving chaos: how distributed tracing brings observability to a service mess