microservices-patterns

Microservices decomposition uses the Strangler Fig pattern to incrementally extract bounded contexts identified via business capabilities or DDD, ensuring the monolith remains functional during migration. This approach defines clear service boundaries and contracts to maintain system integrity throughout the transition.

The Saga pattern manages distributed data consistency through compensating workflows, avoiding global transactions across services. You need Saga-based eventual consistency when applying database-per-service ownership, ensuring microservices remain decoupled while coordinating multi-step business processes safely.

A circuit breaker improves microservices resilience by stopping cascading failures when a downstream service is unavailable. Combined with retry strategies using exponential backoff and bulkheads for resource isolation, it prevents total system collapse during partial outages and maintains service availability.

Synchronous request/response via REST, gRPC, or GraphQL suits operations needing immediate consistency, while asynchronous event-driven messaging via queues or streaming enables loose coupling and scalability. Your choice depends on whether the interaction requires real-time responses or can tolerate eventual consistency.

Defining service boundaries relies on identifying bounded contexts through business capabilities or domain-driven design. In event-driven architectures, you establish explicit message and event contracts alongside strict data ownership rules, ensuring each service maintains autonomy while communicating asynchronously.

Database-per-service prevents tight coupling by giving each microservice exclusive ownership of its data. This constraint eliminates global transactions, requiring you to adopt Saga-based compensating workflows and event-driven coordination to maintain data consistency across service boundaries safely.