Hiding The Lead
summary
Argues that independent deployability depends on strict information hiding and shows concrete failure modes: shared databases create common coupling and distributed monoliths, direct data access creates content coupling (pathological), while inevitable interactions are domain coupling. Recommends owning aggregates (orders state machine), exposing minimal stable APIs, using explicit schemas like OpenAPI JSON Schema Protocol Buffers and running compatibility checks (openapi-diff style) to catch breaking changes before deploy, reducing outages and restoring the ability to change services safely.
AI-written summary · about
chapters (8)
- 0:00Conference welcome
- 1:08Introduction and microservices overview
- 3:57Independent deployability in microservices
- 13:01Information hiding in microservices
- 18:37Coupling and cohesion fundamentals
- 23:54Types of coupling in microservices
- 33:04Summary and key takeaways
- 34:29Q&A and closing remarks
