Chapter 1. Performance Hunting Tips

Solving performance problems without foresight and planning is painful. You waste time and are constantly frustrated as the cause of the problem repeatedly slips through your fingers. By following the right set of procedures, you can transform a frustrating performance hunt into an interesting detective story. Each piece of information leads you closer to the true culprit. People can not always be trusted. The evidence will be your only friend. As you investigate the problem, it will take unusual twists and turns, and the information that you discovered in the beginning of the hunt may turn out to be what helps you figure out the problem in the end. The best part is that you will have a thrill of adrenaline and sense of accomplishment when you finally nab the "bad guy" and fix the problem.

If you have never investigated a performance problem, the first steps can be overwhelming. However, by following a few obvious and nonobvious tips, you can save time and be well on your way to finding the cause of a performance problem. The goal of this chapter is to provide you with a series of tips and guidelines to help you hunt a performance problem. These tips show you how to avoid some of the common traps when investigating what is wrong with your system or application. Most of these tips were hard-learned lessons that resulted from wasted time and frustrating dead ends. These tips help you solve your performance problem quickly and efficiently.

After reading this chapter, you should be able to

• Avoid repeating the work of others.

• Avoid repeating your own work.

• Avoid false leads that result from gathering misleading information.

• Create a useful reference document about your investigation.

Although no performance investigation is flawless (you will almost always say, "If only I would have thought of that first"), these tips help you to avoid some of the common mistakes of a performance investigation.