Performance problems in Drupal are like cancer. Its best to prevent them by following best practices and regular testing. In this session we’ll simulate a nightmare: A poorly optimized site that’s being brought down by traffic spikes.

We’ll begin by purposefully taking down a real site. During this time we’ll cover:
1. The basics of server monitoring - what you never want to see in process lists.
2. How to resurrect a dead site.
3. Some quick stop-gap fixes to use in emergencies.

We’ll then do a post-mortem of the site’s death. You’ll learn:
1. How to parse slow query logs.
2. How XHProf reveals the worst offenders to you site’s performance.
3. Basic techniques for fixing bottle necks.
4. How to easily benchmark your codebase.
5. The most common performance problems you’re likely to see in the real world.