Observability has been defined as making it possible in the future to ask questions about the functioning of our code without having to decide on those questions in advance. For testing and build engineering, this represents, as it does with production services, a change in how we instrument and understand our systems. CI pipelines are often designed to throw away all information about the CI process except for what we have explicitly chosen to keep. I will talk about how we can make observability of the pipeline a reality, and how central message brokers fit into that design.