There are Table 19.1.
Page timings can be useful in determining the effect of either page coding or underlying database operations and caching. In the reports, this is primarily determined by looking at the TTFB (Time To First Byte) and TTLB (Time To Last Byte).
Noting these times and making changes to the logic or database access gives you a good way to track the resulting changes and make sure that you haven't introduced a delay you didn't mean to!
The performance counters for this testing, and many system problems as well, provide a wealth of information. The performance counters enable you to find when your clients and/or servers have reached their peaks in terms of processor, memory, disk I/O, and network throughput.
The idea is that you should make sure that the testing isn't bottlenecked by the client computer that's driving the testing. If the client computer CPU usage is above 90%, it's probably the limiting factor in being able to produce a load on the Web server.
By looking at various counters on the Web server, you should be able to determine performance issues. A rule of thumb is that the Web server CPU usage should be at least 80%. At that point, you'll see the maximum limit of the Web server and not the limitations of the client test computer.
Comparing RPS on various test runs will help you tune the Web server and the Web application.