How renanssantos1/arquitetura-de-microsservicos-padrao-saga-orquestrado works
Project type and purpose
• This is a full-stack, microservices-based application implementing the Saga Orchestrator pattern.
• It consists of five Spring Boot services that coordinate a distributed transaction via Apache Kafka.
• Datastores: MongoDB (order history), PostgreSQL (product validation, payment, inventory). No persistent store for the orchestrator.
File structure
• Top-level directories:
• order-service
• orchestrator-service
• product-validation-service
• payment-service
• inventory-service
• build.py, docker-compose.yml, README.md
• Each service is a standalone Spring Boot/Gradle project with its own src/main/java and application.yml.
• Configurations for Kafka, exception handlers, DTOs, producers/consumers are under each service’s config and core packages.
• Docker artifacts: Dockerfile in each service, a single docker-compose.yml at root, and a Python build script.
README for architectural insights
• Five microservices
• An orchestrator that drives saga steps
• Kafka as message broker
• Three PostgreSQL DBs and one MongoDB
• Endpoints:
• Order-Service exposes REST to start/retrieve saga data (Swagger UI on port 3000).
• Execution via docker-compose or build.py.
a. Main components
• Order-Service (MongoDB)
• Orchestrator-Service (no DB)
• Product-Validation-Service (PostgreSQL)
• Payment-Service (PostgreSQL)
• Inventory-Service (PostgreSQL)
• Apache Kafka cluster (topics for commands & events)
• Databases: 3× PostgreSQL instances, 1× MongoDB
• Redpanda Console (optional visibility)
• docker-compose orchestration layer
b. Relationships and interactions
1. Client → Order-Service (REST API “start saga”)
2. Order-Service produces “OrderCreated” event to Kafka topic
3. Orchestrator-Service consumes “OrderCreated,” logs saga state, and publishes “ValidateProduct” command
4. Product-Validation-Service consumes “ValidateProduct,” writes to its DB, then publishes “ProductValidated” event
5. Orchestrator consumes “ProductValidated,” logs state, sends “ProcessPayment” command
6. Payment-Service consumes “ProcessPayment,” writes to its DB, publishes “PaymentProcessed” event
7. Orchestrator consumes “PaymentProcessed,” logs state, sends “AllocateInventory” command
8. Inventory-Service consumes “AllocateInventory,” writes to its DB, publishes “InventoryAllocated” event
9. Orchestrator consumes “InventoryAllocated,” marks saga as complete; final status can be retrieved via Order-Service
c. Architectural patterns and principles
• Saga Orchestrator pattern for distributed transactions
• Event-driven, asynchronous communication via Kafka topics
• Single responsibility per microservice
• Polyglot persistence (MongoDB for events/logs, PostgreSQL for domain data)
• Centralized error handling and DTO mapping in each service
d. Key technologies and libraries
• Java 17, Spring Boot 3, Spring Kafka
• Docker & docker-compose
• Kafka (Redpanda Console)
• Gradle build scripts (automated via build.py)
• MongoDB, PostgreSQL
Important details
• Error or compensation flows (if implemented)
• Orchestration logic outside of any database