The performance loop—A practical guide to profiling and benchmarking
summary
Shows a practical pipeline: a console test harness copies the messaging pipeline into a repo, removes DI and IO noise, and runs profilers (JetBrains dotTrace/dotMemory plus perfu) and flame graphs to find hot paths. Uses Benchmark.net for controlled permutations and compares before/after with a resolve-compare flow. Achieves concrete wins: removes ~20–27 MB allocations, cuts infrastructure CPU by ~32.3% on publish and ~14% on receive, and reports 10x faster execution and 5x faster pipeline in focused benchmarks; cites a 24% monthly cost reduction migration number.
AI-written summary · about
chapters (11)
- 0:00Introduction
- 5:43Performance optimizations matter
- 11:48Performance optimization loop
- 21:13Profiling the pipeline
- 27:36Memory hotspots
- 32:54CPU flame graphs
- 37:52Reliable benchmarks
- 44:12Benchmarking principles
- 53:47Benchmarking exceptions
- 58:01Validating in production
- 1:00:48Summary
related talks
01h 03mPerformance loop—A practical guide to profiling and benchmarking
01h 03mApplication Performance Optimisation in Practice
48mFantastic Performance and Where to Find It
01h 06mGet Better App Performance with this One Weird Trick
57mPragmatic Performance: When to care about perf, and what to do about it
39mProfiling and Benchmarking 101
29mHow to Performance
01h 02mApplication Performance Optimisation in Practice