Why you should never build Microservices - and why we do it anyway
summary
Argues that microservices are worth only when they solve people or scaling problems, not as a default: Queue-it’s virtual waiting room moves from monolith to services to gain per-feature scaling and higher availability but pays for it in CI/CD, distributed tracing, centralized logging and async messaging. A team of twelve developers spends weeks building deployment pipelines and environments; using Docker and AWS Fargate reduces ops pain. Warns of the distributed monolith trap, recommends monolith-first, identify bounded contexts, and extract low-risk services (reporting, identity) when proven.
AI-written summary · about
chapters (36)
- 0:00Introduction
- 1:25What are Microservices
- 3:00Why we like Microservices
- 3:44Layered architecture
- 4:48Complexity
- 5:45Promises
- 6:35Value
- 7:39People
- 8:44Smaller teams
- 11:44Deployment pipelines
- 13:18Advanced Network layer
- 13:57Multiple programming languages
- 14:23Debugging
- 15:41Multiple Services
- 16:15Error Cases
- 16:54Monitoring
- 18:35Tests Debugging
- 19:21Unit Tests
- 20:37Pitfalls
- 21:01Not enough skilled people
- 22:11Distributed monolith
- 25:02Eventually consistency
- 27:07Crossfunctional teams
- 28:16Bounded context domains
- 31:19Availability reliability
- 32:51Thoughts on Microservices
- 34:48Perfection
- 38:44Monolith
- 41:38Picking out pieces
- 44:13Look for bounded contexts
- 44:40Layers
- 45:54Architecture
- 48:59Minimum requirements
- 50:44Development environments
- 52:31Logging monitoring
- 54:12Technology
related talks
55mMicroservices, Where Did It All Go Wrong
42mMicroservices are Not Worth the Trouble!!??
57mEmbracing Microservices
44mMicroservices, Where Did It All Go Wrong
01h 04mDon’t Build a Distributed Monolith
26mPros/Cons of Microservices
53mDon’t Build a Distributed Monolith: How to Avoid Doing Microservices Wrong
01h 02mDesigning Microservices