Performance Measurement Tools

Rarely does a performance issue involve a single item. Usually, sluggish performance is a problem that has been developing over a period of time and might take several cumulative steps to fix. Also, be aware that fixing a bottleneck in one area might degrade the performance of another part of the system. It’s turns into a constant battle that you do not want to let get out of hand. To adequately monitor your system, you’re going to need to focus on three areas:

• Throughput The amount of work that can be performed in a given amount of time. Sometimes we see this measured as TPS (transactions per second) or MIPS (millions of instructions per second).

• Response time The time that a user has to wait for some work to be complete. For example, how long does it take the ls command to execute, or how long before a shell prompt displays on a terminal after the user enters a password?

• Utilization How much of the computer’s resources were used to perform a task.

The following are the system utilities you’ll use to monitor the performance of your system:

• uptime

• ps

• sdtprocess

• prstat

• vmstat

• iostat

• sar

I’ve already described some of these utilities in previous chapters, and the rest I’ll describe in the sections that follow.