Keynote: The Performance Mindset
summary
A clear checklist for avoiding production meltdowns: prepare for the real spike (an example here is 1.4M viewers translating into 200,000 requests per second), generate realistic warmup and scaleable load with tools like Gatling JMeter BlazeMeter, and analyze percentiles (especially the 99th percentile) not just averages. Monitor CPU load versus utilization, run load tests nightly and in CI, visualize metrics, treat performance as continuous work, and optimize perceived responsiveness.
AI-written summary · about
chapters (14)
- 0:00Conference opening
- 1:25Quizzl success and TV disaster
- 6:48Performance as non-functional requirement
- 9:13Outsourcing dilemma and mentorship crisis
- 13:52Eight lessons for performance mindset
- 14:13Lesson 1: Know requirements
- 18:52Lesson two: Prove you meet requirements
- 22:05Lesson 3: Know statistics
- 26:24Lesson 4: Know runtime
- 31:47Lesson 5: Continuous performance work
- 32:52Lesson 6: Integrate performance daily
- 34:17Lesson 7: Next bottleneck expected
- 36:02Lesson 8: Optimize perceived performance
- 39:53Closing thoughts
related talks
28mStress Testing in Production: The New York Times Engineering Survival Guide
41mPerformance Engineering in the Age of DevOps and Agile
57mSite-Speed That Sticks
50mPerformance: Avoiding Pitfalls, Diagnosing Issues & Scaling for Growth
58mCultivating Production Excellence
19mThink Prod Up Front: Shifting Left on Production Excellence
23mHow to tame chaos effectively?
48mHow to not Crash in Prod, Load Testing Crash Course