Your app has a slow endpoint. You’ve added a database index, turned on caching, reached for larger instances — and you’re still guessing. This talk teaches you how to read flamegraphs: the visualization that shows exactly where your code spends time, no guesswork required. We’ll demonstrate the technique through a real production story from GitHub, where profiling a single page cut CPU consumption by 13% and rebuilt developer trust in the merge button on the Pull Request page.

---

Elena Tanasoiu is a Staff Engineer on GitHub’s Performance Engineering team. She spends her time making one of the world’s largest Ruby on Rails applications faster by profiling production traffic, diagnosing bottlenecks, and occasionally convincing teams they don’t need more hardware. She has a fondness for flamegraphs because they never lie. Also enjoys pickles. Emma Gabriel is the Senior Director of Product Management leading GitHub’s frontend platform teams. She spends her time figuring out what people actually mean when they say ‘this feels slow’, and then deciding what to do about it. In a past life, she spent three years as a Ruby on Rails developer; now she has two kids to worry about endlessly, and very little time for anything else.