When building distributed systems, we aim to get a macro view and be able to zoom in when a particular component seems to be at fault for a failure. Luckily, distributed tracing captures the detailed execution of causally-related activities performed by the elements of a distributed system as it processes a given request.

If you wonder how different the execution of the requests from the system’s expected behavior was, join my session to learn how to recognize trace patterns, determine the bottlenecks when a request is too slow, and deal with oversampling or traffic labeling in production.