Microservices – Diptanu Choudhury
summary
Service architecture claims: 500 microservices handle 62 million users and over two billion API requests while deploying several hundred times per week on AWS in active‑active regions. Operational practices: immutable images, cluster units with client‑side load balancing, and heavy monitoring (roughly 30–40% cost). Failure tools and numbers: circuit‑breaking via Hystrix, predictive auto‑scaling with Scryer, and fault injection from the Chaos Monkey family (gorilla, kong) to validate graceful degradation and rollback.
AI-written summary · about
chapters (20)
- 0:00Introduction
- 0:36Agenda
- 1:58Netflix
- 3:31Microservices
- 4:13Netflix on AWS
- 5:51Speed of change and scale
- 7:28Network failures
- 10:07Fail gracefully
- 11:17Tame complexity at scale
- 16:04Dependencies
- 16:22Dependencies as Microservices
- 17:00cascading failures
- 17:20hystrix
- 18:13reactive scaling
- 18:59predictive scaling
- 19:30failure recovery
- 19:45Chaos Monkey
- 20:18Chaos Gorilla
- 20:39Chaos Kong
- 21:32Design to embrace failures
related talks
01h 00mSix Secrets to Succeeding with Microservices
31mPast, Present and Future of Microservices
26mFighting the Good Fight at the Hot Gates of Micro-Services
01h 01mMicroservices Gone Wrong
49mChallenges in Implementing Microservices
01h 03mMicroservices: Decomposing Applications for Testability and Deployability
24mMicroservices in Practice
52mWhat Does a Good Microservice Architecture Look Like?