Spring Boot 开发模式

用于可扩展、生产级服务的 Spring Boot 架构和 API 模式。

何时激活

• 使用 Spring MVC 或 WebFlux 构建 REST API
• 构建控制器 → 服务 → 仓库层结构
• 配置 Spring Data JPA、缓存或异步处理
• 添加验证、异常处理或分页
• 为开发/预发布/生产环境设置配置文件
• 使用 Spring Events 或 Kafka 实现事件驱动模式

REST API 结构

java
@RestController
@RequestMapping("/api/markets")
@Validated
class MarketController {
  private final MarketService marketService;

  MarketController(MarketService marketService) {
    this.marketService = marketService;
  }

  @GetMapping
  ResponseEntity<Page<MarketResponse>> list(
      @RequestParam(defaultValue = "0") int page,
      @RequestParam(defaultValue = "20") int size) {
    Page<Market> markets = marketService.list(PageRequest.of(page, size));
    return ResponseEntity.ok(markets.map(MarketResponse::from));
  }

  @PostMapping
  ResponseEntity<MarketResponse> create(@Valid @RequestBody CreateMarketRequest request) {
    Market market = marketService.create(request);
    return ResponseEntity.status(HttpStatus.CREATED).body(MarketResponse.from(market));
  }
}

仓库模式 (Spring Data JPA)

java
public interface MarketRepository extends JpaRepository<MarketEntity, Long> {
  @Query("select m from MarketEntity m where m.status = :status order by m.volume desc")
  List<MarketEntity> findActive(@Param("status") MarketStatus status, Pageable pageable);
}

带事务的服务层

java
@Service
public class MarketService {
  private final MarketRepository repo;

  public MarketService(MarketRepository repo) {
    this.repo = repo;
  }

  @Transactional
  public Market create(CreateMarketRequest request) {
    MarketEntity entity = MarketEntity.from(request);
    MarketEntity saved = repo.save(entity);
    return Market.from(saved);
  }
}

DTO 和验证

java
public record CreateMarketRequest(
    @NotBlank @Size(max = 200) String name,
    @NotBlank @Size(max = 2000) String description,
    @NotNull @FutureOrPresent Instant endDate,
    @NotEmpty List<@NotBlank String> categories) {}

public record MarketResponse(Long id, String name, MarketStatus status) {
  static MarketResponse from(Market market) {
    return new MarketResponse(market.id(), market.name(), market.status());
  }
}

异常处理

java
@ControllerAdvice
class GlobalExceptionHandler {
  @ExceptionHandler(MethodArgumentNotValidException.class)
  ResponseEntity<ApiError> handleValidation(MethodArgumentNotValidException ex) {
    String message = ex.getBindingResult().getFieldErrors().stream()
        .map(e -> e.getField() + ": " + e.getDefaultMessage())
        .collect(Collectors.joining(", "));
    return ResponseEntity.badRequest().body(ApiError.validation(message));
  }

  @ExceptionHandler(AccessDeniedException.class)
  ResponseEntity<ApiError> handleAccessDenied() {
    return ResponseEntity.status(HttpStatus.FORBIDDEN).body(ApiError.of("Forbidden"));
  }

  @ExceptionHandler(Exception.class)
  ResponseEntity<ApiError> handleGeneric(Exception ex) {
    // Log unexpected errors with stack traces
    return ResponseEntity.status(HttpStatus.INTERNAL_SERVER_ERROR)
        .body(ApiError.of("Internal server error"));
  }
}

缓存

需要在配置类上使用 @EnableCaching。

java
@Service
public class MarketCacheService {
  private final MarketRepository repo;

  public MarketCacheService(MarketRepository repo) {
    this.repo = repo;
  }

  @Cacheable(value = "market", key = "#id")
  public Market getById(Long id) {
    return repo.findById(id)
        .map(Market::from)
        .orElseThrow(() -> new EntityNotFoundException("Market not found"));
  }

  @CacheEvict(value = "market", key = "#id")
  public void evict(Long id) {}
}

异步处理

需要在配置类上使用 @EnableAsync。

java
@Service
public class NotificationService {
  @Async
  public CompletableFuture<Void> sendAsync(Notification notification) {
    // send email/SMS
    return CompletableFuture.completedFuture(null);
  }
}

日志记录 (SLF4J)

java
@Service
public class ReportService {
  private static final Logger log = LoggerFactory.getLogger(ReportService.class);

  public Report generate(Long marketId) {
    log.info("generate_report marketId={}", marketId);
    try {
      // logic
    } catch (Exception ex) {
      log.error("generate_report_failed marketId={}", marketId, ex);
      throw ex;
    }
    return new Report();
  }
}

中间件 / 过滤器

java
@Component
public class RequestLoggingFilter extends OncePerRequestFilter {
  private static final Logger log = LoggerFactory.getLogger(RequestLoggingFilter.class);

  @Override
  protected void doFilterInternal(HttpServletRequest request, HttpServletResponse response,
      FilterChain filterChain) throws ServletException, IOException {
    long start = System.currentTimeMillis();
    try {
      filterChain.doFilter(request, response);
    } finally {
      long duration = System.currentTimeMillis() - start;
      log.info("req method={} uri={} status={} durationMs={}",
          request.getMethod(), request.getRequestURI(), response.getStatus(), duration);
    }
  }
}

分页和排序

java
PageRequest page = PageRequest.of(pageNumber, pageSize, Sort.by("createdAt").descending());
Page<Market> results = marketService.list(page);

容错的外部调用

java
public <T> T withRetry(Supplier<T> supplier, int maxRetries) {
  int attempts = 0;
  while (true) {
    try {
      return supplier.get();
    } catch (Exception ex) {
      attempts++;
      if (attempts >= maxRetries) {
        throw ex;
      }
      try {
        Thread.sleep((long) Math.pow(2, attempts) * 100L);
      } catch (InterruptedException ie) {
        Thread.currentThread().interrupt();
        throw ex;
      }
    }
  }
}

速率限制 (过滤器 + Bucket4j)

安全须知：默认情况下 X-Forwarded-For 头是不可信的，因为客户端可以伪造它。 仅在以下情况下使用转发头：

1. 您的应用程序位于可信的反向代理（nginx、AWS ALB 等）之后
2. 您已将 ForwardedHeaderFilter 注册为 bean
3. 您已在应用属性中配置了 server.forward-headers-strategy=NATIVE 或 FRAMEWORK
4. 您的代理配置为覆盖（而非追加）X-Forwarded-For 头

当 ForwardedHeaderFilter 被正确配置时，request.getRemoteAddr() 将自动从转发的头中返回正确的客户端 IP。 没有此配置时，请直接使用 request.getRemoteAddr()——它返回的是直接连接的 IP，这是唯一可信的值。

java
@Component
public class RateLimitFilter extends OncePerRequestFilter {
  private final Map<String, Bucket> buckets = new ConcurrentHashMap<>();

  /*
   * SECURITY: This filter uses request.getRemoteAddr() to identify clients for rate limiting.
   *
   * If your application is behind a reverse proxy (nginx, AWS ALB, etc.), you MUST configure
   * Spring to handle forwarded headers properly for accurate client IP detection:
   *
   * 1. Set server.forward-headers-strategy=NATIVE (for cloud platforms) or FRAMEWORK in
   *    application.properties/yaml
   * 2. If using FRAMEWORK strategy, register ForwardedHeaderFilter:
   *
   *    @Bean
   *    ForwardedHeaderFilter forwardedHeaderFilter() {
   *        return new ForwardedHeaderFilter();
   *    }
   *
   * 3. Ensure your proxy overwrites (not appends) the X-Forwarded-For header to prevent spoofing
   * 4. Configure server.tomcat.remoteip.trusted-proxies or equivalent for your container
   *
   * Without this configuration, request.getRemoteAddr() returns the proxy IP, not the client IP.
   * Do NOT read X-Forwarded-For directly—it is trivially spoofable without trusted proxy handling.
   */
  @Override
  protected void doFilterInternal(HttpServletRequest request, HttpServletResponse response,
      FilterChain filterChain) throws ServletException, IOException {
    // Use getRemoteAddr() which returns the correct client IP when ForwardedHeaderFilter
    // is configured, or the direct connection IP otherwise. Never trust X-Forwarded-For
    // headers directly without proper proxy configuration.
    String clientIp = request.getRemoteAddr();

    Bucket bucket = buckets.computeIfAbsent(clientIp,
        k -> Bucket.builder()
            .addLimit(Bandwidth.classic(100, Refill.greedy(100, Duration.ofMinutes(1))))
            .build());

    if (bucket.tryConsume(1)) {
      filterChain.doFilter(request, response);
    } else {
      response.setStatus(HttpStatus.TOO_MANY_REQUESTS.value());
    }
  }
}

后台作业

使用 Spring 的 @Scheduled 或与队列（如 Kafka、SQS、RabbitMQ）集成。保持处理程序是幂等的和可观察的。

可观测性

• 通过 Logback 编码器进行结构化日志记录 (JSON)
• 指标：Micrometer + Prometheus/OTel
• 追踪：带有 OpenTelemetry 或 Brave 后端的 Micrometer Tracing

生产环境默认设置

• 优先使用构造函数注入，避免字段注入
• 启用 spring.mvc.problemdetails.enabled=true 以获得 RFC 7807 错误 (Spring Boot 3+)
• 根据工作负载配置 HikariCP 连接池大小，设置超时
• 对查询使用 @Transactional(readOnly = true)
• 在适当的地方通过 @NonNull 和 Optional 强制执行空值安全

记住：保持控制器精简、服务专注、仓库简单，并集中处理错误。为可维护性和可测试性进行优化。