Is it Time to Abandon Microservices? Modular Monoliths Rescue Your Wallet!
summary
Argues that microservices often cost more because cloud and observability overheads grow with scale and pay the big vendors, that the Saga pattern forces manual compensating transactions and makes debugging and eventual consistency painful, and that a modular monolith built with Spring Modulith enforces module boundaries at test time, emits UML dependency graphs, keeps events transactional, lets teams cut and paste modules into true services later, and that level two caching can remove roughly 80–90% of read IO.
AI-written summary · about
chapters (16)
- 0:00Introduction and speaker background
- 1:22Microservices and monoliths history
- 4:52Misconceptions about microservices
- 8:21Practical challenges of microservices
- 13:42Database design in microservices
- 18:23Why microservices are complex
- 21:50Introducing modular monoliths
- 26:55Scaling modular monoliths
- 31:11Language choice and microservices
- 33:41Team structure and microservices
- 38:44Observability and failure handling
- 41:55Cost comparison monolith vs microservices
- 43:41Summary and future recommendations
- 48:01Q&A: Refactoring to microservices
- 49:50Q&A: Cultural and hiring challenges
- 51:56Q&A: Database refactoring difficulties
