How We Engineered Value in Times of COVID?
summary
Retail platform scales from 4x to 50x demand by specific actions: migrate Cloud SQL → MySQL 8.0 and Percona cluster with ProxySQL, move read traffic to replicas and refactor hot queries (reduce select for update locks), and implement sharding with Kafka-based EDA for H0/H1..HN clusters plus cache invalidation. Teams build a production traffic mirroring and a blue/green analyzer dashboard; rollout and sharding tests complete in ~1–2 weeks and succeed in production traffic spikes.
AI-written summary · about
chapters (14)
- 0:00Introduction and context
- 1:05COVID-19 lockdown challenges
- 3:15Traffic surge and team coordination
- 5:00Database scalability issues
- 7:50MySQL upgrades and optimizations
- 11:2250x traffic prediction and sharding
- 14:10Challenges in sharding implementation
- 16:16Production traffic mirroring framework
- 18:43Blue-green deployment analyzer
- 19:24Fulfillment and MMFCs
- 21:48Breaking self-imposed limits
- 22:53Customer value and business growth
- 25:21Call to action for COVID-19 support
- 28:04Q&A and community initiatives
related talks
44mIntroduction to application scaling
29mEvolve your web application for scale
50mBuilding and Operating Online Shops under Extreme Load
46mScaling Pinterest
26mHow We Prepped Our Infrastructure for a Super-Bowl Sized Traffic
32mDiscover the Modern Way to Build Real-Time Analytics Solutions at Any Scale
35mZero-downtime deployment on Kubernetes
24mHow to design your team for scale