The last technique we'll discuss in this section is meant for distributed systems. When looking at the overall system, often deployed in the cloud, profiling your software on one box won't tell you the whole story. In such a scope, your best bet would be to trace the requests and responses flowing through your system.
Tracing is a way to log the execution of your code. It's often used when a request (and sometimes its response) has to flow through many parts of your system. Usually, such messages are being traced along the route, with timestamps being added at interesting points of execution.
