Microservice architecture organizes independently deployable services around separate capabilities. A customer-profile service can validate and store profile changes, then publish events to Kafka; another service can combine profile, billing, and interaction events into a richer view. section 3.1
Kafka provides an event-driven integration layer between services, reducing direct coupling while supporting scalability and fault tolerance. Each service owns the code that produces, consumes, validates, processes, and acknowledges its messages. section 3.1 section 5.5
Design decisions balance development simplicity with operational control:
- Use Spring-Kafka or Micronaut-Kafka to simplify client development, accepting that frameworks can obscure Kafka internals. section 3.1
- Use low-level Kafka clients when custom message conversion, dead-letter routing, consumer concurrency, or offset handling requires precise control. section 3.1
- Choose Kafka according to non-functional requirements; RabbitMQ or Apache Pulsar may better fit low-latency transactional messaging or other specialized needs. section 5.5
