To manage complex distributed systems, you need to be able to observe and understand what’s happening to all of the components that make up the system. Observability, however, hinges on the assumption that every component can generate information about what’s happening with it, and in an event-driven system, that can be quite complicated. 

For example, if you have an application that executes activities A, B, and C and publishes a message to an event broker, which then goes to a queue, you would want to know what happened from start to finish: from the publishing application, to and within the broker, all the way to the receiving application, for every transactional event.