You’ve built some software that users love, and while its use starts to become more frequent it also becomes more critical. Suddenly the failures, inevitable and previously merely a nuisance, have become critical and emotionally fraught as your users see reputation, revenue, and even job security imperiled!
Bugs come in many flavors. None quicken the pulse as much as a critical system suffering from poor performance. They can be hard to drive to root cause, hard to fix, hard to reproduce, and even hard for the user to describe beyond “it’s slow (fix it!)”. Fortunately we’re not completely alone in those lonely moments: we have the lessons and—critically—the tools from the engineers who have survived similar situations.
DTrace is one such tool, born from experiences debugging functional and performance problems on the world’s most critical systems for over a decade. In this talk we will walk through the basics of DTrace and then focus on a case study of using it to solve a performance problem while turning an audibly angry customer into a happy one.
Tags: infrastructure, software development process
Location: Salon C
April 12th, 2016
11:30 AM - 12:30 PM