Avoiding Microservice Megadisasters
summary
A rewrite of a monolith into many tiny services takes 18 months and hundreds of engineers and yields a site that takes 9.5 minutes to load because every page triggers hundreds of chained API calls with no fault tolerance. The fix is to give subsystems ownership: build a search service that owns a denormalized index, pull data via scheduled exports and DB triggers into Microsoft FAST → elastic search, and serve results from one DB so the response drops to sub‑second and failure surface shrinks.
AI-written summary · about
chapters (20)
- 0:00Intro
- 0:24Avoiding Microservice Megadisasters
- 2:01Dellenovacer Computers
- 8:52Netflix-oriented architecture
- 9:38Core principles
- 11:43Pre-production go-live
- 17:20API hell
- 19:19Doing the math
- 22:24How Not to Do It (MS Style)
- 23:58Traditional n-tier
- 24:24Areas of Responsibility
- 26:09what is a Service? Software that is...
- 27:43what is a Microservice?
- 30:43what is a Bounded context?
- 35:59Grounding in reality
- 41:16Data ownership
- 43:04Catalog
- 44:49Pricing
- 45:25Localization
- 46:39Content
related talks
42mAvoiding Microservice Megadisaster
45mAvoiding Microservice Megadisasters
42mDo proper monoliths before you do microservices!
36mDo Proper Monoliths before You Do Microservices!
48mMicroMonolith
54mFrom a distributed monolith to a microservices solution
43mMicroMonolith - Top anti-patterns of distributed systems
43mGo micro with Microservices