How do you test your failure scenarios explicitly across Service APIs?. How can you take control of writing Automated Integration tests that simulate failure cases and test your resiliency and fallbacks?

As we design and build micro services exposing APIs, it becomes very critical to think of all the failure points that can happen and how they will manifest on the response. Its also important that as we plan automated Integration Testing on the APIs that run in CI/CD pipelines to be able to represent both success and failure test cases. Existing approaches to do so have been hard and usually need the whole environment in a failure more to be able to simulate or mock error cases from downstream dependencies.

In this session, Rajat will introduce AutoFIT, a light-weight framework he co-authored with his colleague Arun Lingala, which helps introduce failures using an expression in request header. The framework when consumed in the Service can intercept the expression and simulate a downstream dependency to return a desired error instead and help the code path enter the failure mode. This gives complete control on simulating a desired error for a request without putting the entire system in a failure mode.