High Performance MySQL, 3rd Edition

Chapter 14. Application-Level Optimization

If you spend a lot of time improving MySQL’s performance, it’s easy to get tunnel vision and forget to focus on the user’s experience. You may step back for a bit and realize that MySQL is so highly optimized that it’s contributing only a tiny fraction of the response time the user sees, and it’s time to focus elsewhere. This is a great insight (especially for a DBA), and it’s exactly the right thing to do. But what is causing problems, if not MySQL? The answer can be found most reliably and quickly by measuring, using the techniques we showed in Chapter 3. If your profiling is thorough and you follow a logical process, it should not be hard to find the source of your problem. Sometimes, though, even when the problem is MySQL, it might be easiest to solve it in another part of the system!

No matter where the problem lies, there’s sure to be at least one great tool available to help you measure it, often for free. For example, if you have issues with JavaScript or page rendering, you can use the profiler included with the Firebug extension for the Firefox web browser, or you can use the YSlow tool from Yahoo!. We mentioned several application-level tools in Chapter 3. Some tools even profile the whole stack; New Relic is an example of a tool that profiles the frontend, application, and backend of web applications.

Common Problems

We see the same problems over and over again in applications, often because people have used poorly designed off-the-shelf ...