From Monolith to (µ)Services. What happens with my Data?
summary
Concrete recipe to migrate a monolith piece-by-piece: use the Strangler fig pattern and change-data-capture so MySQL transaction logs (via Debezium) publish JSON events to Kafka, Kafka Connect sinks replicate to MongoDB, and NGINX Ingress routes reads first then writes to the new Quarkus owner service. Demo runs on Kubernetes with Strimzi managing Kafka, supports snapshotting to seed existing rows, lets reads be switched over, then writes, and permits instant rollback; expect eventual consistency windows and connector startup delays.
AI-written summary · about
chapters (8)
- 0:00Start
- 4:43Baby Steps (safe)
- 5:25Application
- 7:52Dual Writes
- 9:35Change Data Capture
- 11:23What's Kafka?
- 16:40Data Change Events
- 18:54Strangler Fig Pattern
related talks
47mFrom Monolith to Microservices With the Strangler Pattern
24mModernise your monolith using strangler pattern
01h 05mMigrating to Microservice Databases: From Relational Monolith to Distributed Data
31mMonolith to Microservices: A Guide to Seamless Transitions
42mQ&A - Monolith to Microservices
55mThe hardest part of microservices: Your Data
48mFrom Monolith to Micro-services with Kubernetes
47mThree Microservice Patterns to Tear Down Your Monoliths