To troubleshoot a devious performance issue or understand the performance characteristics of an application, developers like to use profilers. But what do the numbers they produce actually mean? Are they accurate? How does a profiler actually measure them? What are potential sources of errors? Can you compare those numbers with coworkers’? Is it possible to do anything useful with the produced results? This session provides answers to these questions. You will learn how to judge results produced by profilers, find out about the basic measurement techniques of profilers, and learn how to configure them correctly. The presentation shows many problems in live coding examples. Everything mentioned in the session is applicable to any tool.