How can we build data pipelines that are resilient to change? Data usually outlives application code, and we have to be prepared to deploy streams that can cope with the evolution of that data that is in motion. This talk will discuss the approach and supporting patterns to write resilient data microservices with Spring Cloud Stream and Spring Cloud Dataflow. We will discuss the role of a centralized Schema repository and how can we work with different data models and protocols to achieve schema evolution. A centralized schema service will also help us understand what kind of data is currently being served and consumed by the different parts of our distributed system.