Chopping the monolith
summary
Argues that only hot-changing capabilities need independent deploys and proposes a middle path: keep stable core and move fast-changing parts to FaaS. Shows a concrete migration of a pricing module from a Java monolith into Azure Functions, fronted by Apache APISIX configured via etcd, with Docker Compose and a SQL backend. Notes provenance in rule engines, claims serverless reduces lead time, and reports the live demo routing to Azure failed once but worked in prior runs.
AI-written summary · about
chapters (8)
- 0:00Introduction
- 1:11Microservices hype reality
- 4:27Conway’s law and teams
- 9:47Why microservices fail
- 15:02Legacy deployment struggles
- 21:08Strangling the monolith
- 22:36Live demo setup
- 32:01Q&A and resources
related talks
35mChopping the monolith
31mChopping the monolith
37mChopping the monolith
42mChopping the Monolith
42mChopping the monolith
45mThe Evolution From Monolith to Microservices to Serverless and Beyond
42mHelp! I've created a serverless monolith!
33mFireside chat - Monoliths, Microservices & Modernization – Practical Strategies
