Fighting the Good Fight at the Hot Gates of Micro-Services
summary
The site runs roughly 300 microservices that produce a daily noon traffic spike and the team moves features by classifying services, assigning departments of ~16–20 engineers, and tracking ownership so unknowns fall from ~33% to ~18%. The stack shifts from Ruby to Java to Scala for performance, clients become thin clients generated from API docs to avoid dependency hell, and deployments consolidate onto AWS with canary releases and a small Python layer that proves safe testing in production.
AI-written summary · about
chapters (21)
- 0:00Intro
- 2:42Selling every day at noon
- 5:00The Hype Cycle
- 6:55Staging Environments
- 8:30Lessons from the Slope
- 9:50Emergent Architecture
- 10:24Spreadsheets
- 11:15Classification
- 13:00Pivot Table
- 14:25Stop Making Decisions
- 15:07Ownership
- 16:25Architecture Board
- 18:20Ideal Team Size
- 19:08Time Management
- 19:41Risk Management
- 20:33Tracking Ownership
- 22:17Consumer Dependencies
- 22:50Inversion
- 23:27Zero dependency
- 23:45Building Snowflakes
- 24:22Deployment Pipeline
related talks
23mMicroservices – Diptanu Choudhury
23mMicroservices – Sean Smith
18mMicroservices Out In the Wild
42mInnovating Faster with Micro-Service Architecture...
35mMicroservices: A Tale of 3 Heroes
56mFrom Homogeneous Monolith to Heterogeneous Microservices Architecture
01h 07mMicroservices from the Trenches. A Guide through the Architecture, Design, Code
33mAdventures with Micro Services in Rails