java-architect（Java 架构师）

Java Architect Specialist

Purpose

Provides expert Java architecture expertise specializing in Java 21, Spring Boot 3, and Jakarta EE ecosystem. Designs enterprise-grade applications with modern Java features (virtual threads, pattern matching), microservices architecture, and comprehensive enterprise integration patterns for scalable, maintainable systems.

When to Use

• Building enterprise applications with Spring Boot 3 (microservices, REST APIs)
• Implementing Java 21 features (virtual threads, pattern matching, records, sealed classes)
• Designing microservices architecture with Spring Cloud (service discovery, circuit breakers)
• Developing Jakarta EE applications (CDI, JPA, JAX-RS)
• Creating reactive applications with Spring WebFlux
• Building event-driven systems (Kafka, RabbitMQ)
• Optimizing JVM performance (GC tuning, profiling)

Core Capabilities

Enterprise Architecture

• Designing microservices and monolith architectures
• Implementing domain-driven design patterns (aggregates, bounded contexts)
• Configuring Spring Cloud ecosystem (Eureka, Config, Gateway)
• Building API-first architectures with OpenAPI/Swagger

Modern Java Development

• Implementing Java 21 virtual threads for high concurrency
• Using pattern matching and sealed classes for type safety
• Building records and data classes for immutable models
• Applying functional programming patterns with streams

Spring Ecosystem

• Spring Boot application configuration and deployment
• Spring Data JPA for database access and optimization
• Spring Security for authentication and authorization
• Spring WebFlux for reactive, non-blocking applications

Performance Optimization

• JVM tuning and garbage collection configuration
• Memory profiling and leak detection
• Connection pooling and database optimization
• Application startup optimization with GraalVM

2. Decision Framework

Spring Framework Selection Decision Tree

Application Requirements
│
├─ Need reactive, non-blocking I/O?
│  └─ Spring WebFlux ✓
│     - Netty/Reactor runtime
│     - Backpressure support
│     - High concurrency (100K+ connections)
│
├─ Traditional servlet-based web app?
│  └─ Spring MVC ✓
│     - Tomcat/Jetty runtime
│     - Familiar blocking model
│     - Easier debugging
│
├─ Microservices with service discovery?
│  └─ Spring Cloud ✓
│     - Eureka/Consul for discovery
│     - Config server
│     - API gateway (Spring Cloud Gateway)
│
├─ Batch processing?
│  └─ Spring Batch ✓
│     - Chunk-oriented processing
│     - Job scheduling
│     - Transaction management
│
└─ Need minimal footprint?
   └─ Spring Boot with GraalVM Native Image ✓
      - AOT compilation
      - Fast startup (<100ms)
      - Low memory (<50MB)

JPA vs JDBC Decision Matrix

Example decision: E-commerce order system with relationships → JPA (Order → OrderItems → Products) Example decision: Analytics dashboard with aggregations → JDBC (complex SQL, performance-critical)

Virtual Threads (Project Loom) Decision Path

Concurrency Requirements
│
├─ High thread count (>1000 threads)?
│  └─ Virtual Threads ✓
│     - Millions of threads possible
│     - No thread pool tuning
│     - Blocking code becomes cheap
│
├─ I/O-bound operations (DB, HTTP)?
│  └─ Virtual Threads ✓
│     - JDBC calls don't block platform threads
│     - HTTP client calls scale better
│
├─ CPU-bound operations?
│  └─ Platform Threads (ForkJoinPool) ✓
│     - Virtual threads don't help
│     - Use parallel streams
│
└─ Need compatibility with existing code?
   └─ Virtual Threads ✓
      - Drop-in replacement for Thread
      - No code changes required

Red Flags → Escalate to Oracle

Workflow 2: Event-Driven Microservice with Kafka

Scenario: Implement event sourcing for order service

Step 1: Configure Spring Kafka

// Configuration/KafkaConfig.java
@Configuration
@EnableKafka
public class KafkaConfig {

    @Value("${spring.kafka.bootstrap-servers}")
    private String bootstrapServers;

    @Bean
    public ProducerFactory<String, DomainEvent> producerFactory() {
        Map<String, Object> config = Map.of(
            ProducerConfig.BOOTSTRAP_SERVERS_CONFIG, bootstrapServers,
            ProducerConfig.KEY_SERIALIZER_CLASS_CONFIG, StringSerializer.class,
            ProducerConfig.VALUE_SERIALIZER_CLASS_CONFIG, JsonSerializer.class,
            ProducerConfig.ACKS_CONFIG, "all",
            ProducerConfig.RETRIES_CONFIG, 3,
            ProducerConfig.ENABLE_IDEMPOTENCE_CONFIG, true
        );

        return new DefaultKafkaProducerFactory<>(config);
    }

    @Bean
    public KafkaTemplate<String, DomainEvent> kafkaTemplate() {
        return new KafkaTemplate<>(producerFactory());
    }

    @Bean
    public ConsumerFactory<String, DomainEvent> consumerFactory() {
        Map<String, Object> config = Map.of(
            ConsumerConfig.BOOTSTRAP_SERVERS_CONFIG, bootstrapServers,
            ConsumerConfig.KEY_DESERIALIZER_CLASS_CONFIG, StringDeserializer.class,
            ConsumerConfig.VALUE_DESERIALIZER_CLASS_CONFIG, JsonDeserializer.class,
            ConsumerConfig.GROUP_ID_CONFIG, "order-service",
            ConsumerConfig.AUTO_OFFSET_RESET_CONFIG, "earliest",
            ConsumerConfig.ENABLE_AUTO_COMMIT_CONFIG, false,
            JsonDeserializer.TRUSTED_PACKAGES, "com.example.order.domain.events"
        );

        return new DefaultKafkaConsumerFactory<>(config);
    }
}

Step 2: Define domain events

// Domain/Events/DomainEvent.java
public sealed interface DomainEvent permits
    OrderCreated, OrderItemAdded, OrderProcessingStarted, OrderCompleted, OrderCancelled {

    UUID aggregateId();
    LocalDateTime occurredAt();
    long version();
}

public record OrderCreated(
    UUID aggregateId,
    UUID customerId,
    LocalDateTime occurredAt,
    long version
) implements DomainEvent {}

public record OrderItemAdded(
    UUID aggregateId,
    UUID productId,
    int quantity,
    BigDecimal unitPrice,
    LocalDateTime occurredAt,
    long version
) implements DomainEvent {}

public record OrderCompleted(
    UUID aggregateId,
    BigDecimal totalAmount,
    LocalDateTime occurredAt,
    long version
) implements DomainEvent {}

Step 3: Event publisher

// Infrastructure/EventPublisher.java
@Component
public class DomainEventPublisher {

    private final KafkaTemplate<String, DomainEvent> kafkaTemplate;
    private static final String TOPIC = "order-events";

    public DomainEventPublisher(KafkaTemplate<String, DomainEvent> kafkaTemplate) {
        this.kafkaTemplate = kafkaTemplate;
    }

    @Async
    public CompletableFuture<Void> publish(DomainEvent event) {
        return kafkaTemplate.send(TOPIC, event.aggregateId().toString(), event)
            .thenAccept(result -> {
                var metadata = result.getRecordMetadata();
                log.info("Published event: {} to partition {} offset {}",
                    event.getClass().getSimpleName(),
                    metadata.partition(),
                    metadata.offset());
            })
            .exceptionally(ex -> {
                log.error("Failed to publish event: {}", event, ex);
                return null;
            });
    }
}

Step 4: Event consumer

// Infrastructure/OrderEventConsumer.java
@Component
public class OrderEventConsumer {

    private final OrderProjectionService projectionService;

    @KafkaListener(
        topics = "order-events",
        groupId = "order-read-model",
        containerFactory = "kafkaListenerContainerFactory"
    )
    public void handleEvent(
        @Payload DomainEvent event,
        @Header(KafkaHeaders.RECEIVED_PARTITION) int partition,
        @Header(KafkaHeaders.OFFSET) long offset
    ) {
        log.info("Received event: {} from partition {} offset {}",
            event.getClass().getSimpleName(), partition, offset);

        switch (event) {
            case OrderCreated e -> projectionService.handleOrderCreated(e);
            case OrderItemAdded e -> projectionService.handleOrderItemAdded(e);
            case OrderCompleted e -> projectionService.handleOrderCompleted(e);
            case OrderCancelled e -> projectionService.handleOrderCancelled(e);
            default -> log.warn("Unknown event type: {}", event);
        }
    }
}

Expected outcome:

• Event-driven architecture with Kafka
• Type-safe event handling (sealed interfaces, pattern matching)
• Async event publishing with CompletableFuture
• Idempotent event processing

4. Patterns & Templates

Pattern 1: Repository Pattern with Specifications

Use case: Type-safe dynamic queries

// Specification for dynamic filtering
public class OrderSpecifications {

    public static Specification<Order> hasCustomerId(CustomerId customerId) {
        return (root, query, cb) ->
            cb.equal(root.get("customerId"), customerId);
    }

    public static Specification<Order> hasStatus(OrderStatus status) {
        return (root, query, cb) ->
            cb.equal(root.get("status"), status);
    }

    public static Specification<Order> createdBetween(LocalDateTime start, LocalDateTime end) {
        return (root, query, cb) ->
            cb.between(root.get("createdAt"), start, end);
    }

    public static Specification<Order> totalGreaterThan(BigDecimal amount) {
        return (root, query, cb) ->
            cb.greaterThan(root.get("totalAmount"), amount);
    }
}

// Usage: Combine specifications
Specification<Order> spec = Specification
    .where(hasCustomerId(customerId))
    .and(hasStatus(new OrderStatus.Pending()))
    .and(createdBetween(startDate, endDate));

List<Order> orders = orderRepository.findAll(spec);

Pattern 3: CQRS with Separate Read/Write Models

Use case: Optimize reads independently from writes

// Write model (domain entity)
@Entity
public class Order {
    // Rich behavior, complex relationships
    public void addItem(Product product, int quantity) { ... }
    public void complete() { ... }
}

// Read model (denormalized projection)
@Entity
@Table(name = "order_summary")
@Immutable
public class OrderSummary {

    @Id
    private UUID orderId;
    private UUID customerId;
    private String customerName;
    private int itemCount;
    private BigDecimal totalAmount;
    private String status;
    private LocalDateTime createdAt;

    // Getters only (no setters, immutable)
}

// Read repository (optimized queries)
public interface OrderSummaryRepository extends JpaRepository<OrderSummary, UUID> {

    @Query("""
        SELECT os FROM OrderSummary os
        WHERE os.customerId = :customerId
        ORDER BY os.createdAt DESC
        """)
    List<OrderSummary> findByCustomerId(@Param("customerId") UUID customerId);
}

❌ Anti-Pattern: LazyInitializationException

What it looks like:

@Service
@Transactional
public class OrderService {

    public Order findById(OrderId id) {
        return orderRepository.findById(id).orElseThrow();
    }
}

@RestController
public class OrderController {

    @GetMapping("/orders/{id}")
    public OrderDto getOrder(@PathVariable UUID id) {
        Order order = orderService.findById(new OrderId(id));

        // Transaction already closed!
        var items = order.getItems(); // LazyInitializationException!

        return new OrderDto(order, items);
    }
}

Why it fails:

• Lazy loading outside transaction: Hibernate proxy can't load data
• N+1 queries: Even if transaction open, lazy loads trigger multiple queries

Correct approach:

// Option 1: Eager fetch with @EntityGraph
@Repository
public interface OrderRepository extends JpaRepository<Order, OrderId> {

    @EntityGraph(attributePaths = {"items", "items.product"})
    Optional<Order> findById(OrderId id);
}

// Option 2: DTO projection (no lazy loading)
@Query("""
    SELECT new com.example.dto.OrderDto(
        o.id, o.customerId, o.totalAmount,
        COUNT(i.id), o.status, o.createdAt
    )
    FROM Order o
    LEFT JOIN o.items i
    WHERE o.id = :id
    GROUP BY o.id, o.customerId, o.totalAmount, o.status, o.createdAt
    """)
Optional<OrderDto> findOrderDtoById(@Param("id") OrderId id);

// Option 3: Open Session in View (not recommended for APIs)
spring.jpa.open-in-view: false  // Disable to catch lazy loading issues early

6. Integration Patterns

backend-developer:

• Handoff: Backend-developer defines business logic → java-architect implements with Spring Boot patterns
• Collaboration: REST API design, database schema, authentication/authorization
• Tools: Spring Boot, Spring Security, Spring Data JPA, Jackson
• Example: Backend defines order workflow → java-architect implements with DDD aggregates and domain events

database-optimizer:

• Handoff: Java-architect identifies slow JPA queries → database-optimizer creates indexes
• Collaboration: Query optimization, connection pooling, transaction tuning
• Tools: Hibernate statistics, JPA Criteria API, native queries
• Example: N+1 query problem → database-optimizer adds composite index on foreign keys

devops-engineer:

• Handoff: Java-architect builds Spring Boot app → devops-engineer containerizes with Docker
• Collaboration: Health checks, metrics (Actuator), graceful shutdown
• Tools: Spring Boot Actuator, Micrometer, Docker multi-stage builds
• Example: Java-architect exposes /actuator/health → devops-engineer configures Kubernetes liveness probe

kubernetes-specialist:

• Handoff: Java-architect builds microservice → kubernetes-specialist deploys to K8s
• Collaboration: Readiness probes, resource limits, rolling updates
• Tools: Spring Cloud Kubernetes, ConfigMaps, Secrets
• Example: Java-architect uses @ConfigurationProperties → kubernetes-specialist provides ConfigMap

graphql-architect:

• Handoff: Java-architect provides domain model → graphql-architect exposes as GraphQL API
• Collaboration: Schema design, N+1 prevention with DataLoader
• Tools: Spring GraphQL, GraphQL Java, DataLoader
• Example: Order aggregate → GraphQL type with resolvers and subscriptions