Today Mark Imbriaco of 37signals posted a short synopsis on their blog Signal Vs Noise of the process he followed to improve the response time of one of their leading products, Basecamp. This is a good example of the kind of continuous tuning and experimentation that dev teams should take on regularly, in addition to adding new features and solving problems which we all get wrapped up in every day.
Mark was able to create an A-B test using new hardware to compare dedicated versus virtualized systems and at the same time compare performance on two different processors. The resulting improvements to Basecamp response time were pretty dramatic. We are proud that Mark makes good use of New Relic RPM to assist him in his diagnosis and solution. Check it out at Signal Vs. Noise.