Microservices for Mortals
summary
Microservices promise faster change but require concrete investments: service discovery and config, centralized logging, metrics, and service discovery and config that actually work in production. Expect to build registries, deploy Docker containers, adopt Hystrix and bulkheads for timeouts and circuit breakers, and choose between REST, binary protocols or messaging like Kafka. Data becomes polyglot persistence and eventual consistency with copy/duplication tradeoffs. Netflix examples scale but only after major tooling, testing, org rework and hard lessons from cascading failures and latency.
AI-written summary · about
chapters (24)
- 0:00Intro
- 0:32About me
- 1:18Italian Food Analogy
- 2:20Pick a side...
- 4:09Where did it come from?
- 10:36What happened to SOA?
- 14:24Which pill do you take?
- 14:46blue pill world
- 18:23There is no spoon
- 19:23Welcome to red pill world
- 20:14About Monoliths
- 21:31When monoliths are bad
- 22:20Monolith to Microservices
- 24:27Initial Investment
- 28:43Data Strategy
- 33:01Synchronous vs Asynchronous
- 33:54Service Communication
- 36:13Avoid re-use traps
- 40:22Conway's Law
- 41:29What it actually means
- 42:56SOA Adoption Model
- 44:53Design for Failure
- 48:42Some take-aways
- 49:44And one more thing...
