Panel Discussion - Performance & Scalability
summary
Company migrates from MySQL 7 to MySQL 8 and hits a blocker on optimized queries against a single 2–3 TB table with 30–40 million rows causing a production delay and a reputation hit despite downtime under one hour, infrastructure grows from ~1,000 to ~3,000 EC2 instances and customers from 200 to 4,000, order processing breaks when Sidekiq jobs scale by CPU instead of queue depth risking seller delivery targets of 85%, teams use Snowflake, Kafka for messaging and LLMs to cut boilerplate and speed onboarding MySQL 7→8 migration caused blocker job autoscaling based on CPU not queue depth
AI-written summary · about
chapters (7)
- 0:00Introduction to the Panel Discussion
- 0:30Meet the Panelists
- 4:17Diving into Performance and Scale
- 4:55Failure Stories and Lessons Learned
- 8:54Monolith vs. Microservices Debate
- 15:28Parallel Workloads and Scalability
- 20:43Engineer Performance and Productivity Tools
related talks
35mSurviving Black Friday: Tales from an e-Commerce Engineer
14mDatabase Performance at Scale for RoR Applications
48mWeb scale for the rest of us
46mScaling With Friends
29mMigrating Shopify’s Core Rails Monolith to Trilogy
27mRails Orbiting the cloud: Mastering the scalability with Kubernetes
28mSimplifying at Scale: 7 Years of Rails Architecture
31mBiggish Data With Rails and Postgresql