Metric Driven Development with Ruby on Rails
summary
Ruby apps get profiled with perf tools.rb, memory with memprof-ish probes and GC counters, and DB issues with Bullet and slow-query watchers. Practical results: a dashboard drops from ~2300ms to ~600ms by stubbing a hotspot (a single method consuming ~60% of runtime), a 70% drop from locating the culprit, and tiny fixes like switching length to count give immediate wins. Targets: ~100ms user-facing server responses or at least under 500ms.
AI-written summary · about
chapters (36)
- 0:00Intro
- 0:23Jeff Casimir
- 1:19Introduction
- 2:09About Jeff
- 3:05Runtime Performance
- 3:53What is Mature
- 5:27Performance is User Experience
- 7:13Goal Setting
- 8:26Performance
- 9:45Performance Perception
- 10:11New Relic
- 10:49Mike Holmes
- 11:09Xrays for Code
- 11:40CPU with Perf Tools
- 12:07Commit Law
- 13:43Initializer
- 14:09Perf Tools
- 14:33Engineers Tools
- 15:26PDF Output
- 16:55Call Chain
- 17:29Sample Project
- 19:32Stack Trace
- 21:14PerfTools
- 22:13CPU
- 23:15Performance Tools
- 23:37CPU Profiling
- 26:18Database Abuse
- 26:36Initializers
- 26:46Queries
- 27:53Query Log
- 28:51Array Length
- 29:31Database Count
- 30:14SlowJam
- 30:56Auto Explain
- 32:05Guessing vs Knowing
- 32:22Outro
