The journey from monolith to microservices
summary
A concrete migration path presents: start with API-level tests and Pact contract testing, reorganize into well-structured monolith modules (use-case objects, mono-repo) with a strict one-table migration rule, then extract a first service (email) using Spring Cloud / Netflix OSS (Eureka for discovery, Hystrix for circuit breakers) and Cloud Foundry for easy ops. The client case cuts outages and scales hotspots but incurs measurable operational cost — extra apps to run, monitor, and the complexity of network partitions.
AI-written summary · about
chapters (31)
- 0:00Introduction
- 0:42About me
- 2:15Background
- 2:52Mike Beranek
- 3:38Simon Brown
- 4:24Netflix and Hulu
- 6:56Engineer tried and true
- 7:36Which comes first
- 9:11Context first
- 9:29Airplane analogy
- 9:56API level tests
- 10:20End state tests
- 10:54Pact
- 11:58Feedback loops
- 12:19Arrange your application
- 12:33Visualise your application
- 12:54What is the domain
- 13:09Maximum number of directories
- 13:23Organization
- 14:43Breaking out components
- 15:24The bonded context
- 15:55Databases
- 16:40Code bases
- 18:37Why microservices
- 20:13Network communication problems
- 21:01Service discovery
- 21:34Monitoring overhead
- 21:57Partition tolerance
- 22:21More network calls
- 25:39Resources
- 27:27Questions
related talks
31mMonolith to Microservices: A Guide to Seamless Transitions
24mPatterns of Evolution from Monolith to Microservices
52mMicroservice Migration Roadmap
57mSplitting Monolith Application to Microservices
32mSplitting the monolith
42mQ&A - Monolith to Microservices
38mFrom Monolith to Microservices
39mConsidering Migrating a Monolith to Microservices?