【技术底稿 15】SpringBoot 异步文件上传实战：多线程池隔离 + 失败重试 + 实时状态推送

一、业务场景在企业级平台中大文件、批量文件上传是高频场景。同步上传极易导致接口超时、前端阻塞、用户体验差等问题。本文基于真实生产实践实现一套通用、高可用、可直接复用的异步文件上传方案异步处理上传逻辑接口快速响应多业务线程池隔离避免互相影响上传失败自动指数退避重试上传结果实时推送前端展示二、核心设计思路使用 SpringBootAsync实现异步上传不阻塞主线程拆分独立线程池文件上传、消息推送业务隔离失败重试采用指数退避策略防止频繁重试压垮服务事务保证文件状态与上传结果一致上传结果通过事件机制推送前端实时感知三、线程池配置业务隔离java运行import org.springframework.context.annotation.Bean; import org.springframework.context.annotation.Configuration; import org.springframework.core.task.TaskExecutor; import org.springframework.scheduling.annotation.EnableAsync; import org.springframework.scheduling.concurrent.ThreadPoolTaskExecutor; import java.util.concurrent.ThreadPoolExecutor; Configuration EnableAsync public class AsyncConfig { /** * 文件上传专用线程池IO 密集型 */ Bean(fileUploadExecutor) public TaskExecutor fileUploadExecutor() { ThreadPoolTaskExecutor executor new ThreadPoolTaskExecutor(); executor.setCorePoolSize(5); executor.setMaxPoolSize(10); executor.setQueueCapacity(100); executor.setThreadNamePrefix(FileUpload-); executor.setRejectedExecutionHandler(new ThreadPoolExecutor.CallerRunsPolicy()); executor.setWaitForTasksCompleteOnShutdown(true); executor.setAwaitTerminationSeconds(60); executor.initialize(); return executor; } /** * 消息推送专用线程池 */ Bean(messagePushExecutor) public TaskExecutor messagePushExecutor() { ThreadPoolTaskExecutor executor new ThreadPoolTaskExecutor(); executor.setCorePoolSize(4); executor.setMaxPoolSize(8); executor.setQueueCapacity(500); executor.setThreadNamePrefix(MsgPush-); executor.setRejectedExecutionHandler(new ThreadPoolExecutor.DiscardOldestPolicy()); executor.initialize(); return executor; } }四、异步上传核心服务java运行import lombok.extern.slf4j.Slf4j; import org.springframework.beans.factory.annotation.Autowired; import org.springframework.scheduling.annotation.Async; import org.springframework.stereotype.Service; import org.springframework.transaction.support.TransactionTemplate; import java.io.File; import java.time.LocalDateTime; Service Slf4j public class AsyncUploadService { Autowired private FileStorageService fileStorageService; Autowired private TransactionTemplate transactionTemplate; Autowired private EventPublisher eventPublisher; /** * 异步上传文件核心方法 */ Async(fileUploadExecutor) public void asyncUpload(Long fileId) { FileResource file fileResourceService.getById(fileId); if (file null) { log.error(文件记录不存在fileId:{}, fileId); return; } try { // 执行上传 String remotePath fileStorageService.upload(file.getTempFilePath()); if (FAILED.equals(remotePath)) { throw new RuntimeException(文件存储服务上传失败); } // 事务更新状态 boolean updateSuccess updateFileStatus(file, remotePath, SUCCESS); if (updateSuccess) { // 删除临时文件 deleteTempFile(file.getTempFilePath()); log.info(文件上传成功fileId:{}, fileId); // 推送成功消息 pushStatusMessage(file, SUCCESS); } } catch (Exception e) { log.error(文件上传异常fileId:{}, fileId, e); handleUploadFailure(file, e); } } /** * 事务内更新文件状态 */ private boolean updateFileStatus(FileResource file, String remotePath, String status) { return transactionTemplate.execute(status - { file.setFilePath(remotePath); file.setUploadStatus(status); file.setUpdateTime(LocalDateTime.now()); return fileResourceService.updateById(file) 0; }); } /** * 删除临时文件 */ private void deleteTempFile(String tempPath) { try { File file new File(tempPath); if (file.exists()) { file.delete(); } } catch (Exception e) { log.warn(临时文件删除失败, e); } } }五、失败重试机制指数退避java运行/** * 上传失败处理 重试 */ private void handleUploadFailure(FileResource file, Exception e) { int currentRetry file.getRetryCount() null ? 0 : file.getRetryCount(); int maxRetry 3; if (currentRetry maxRetry) { // 重试 file.setRetryCount(currentRetry 1); file.setLastRetryTime(LocalDateTime.now()); file.setFailReason(上传失败即将重试 e.getMessage()); fileResourceService.updateById(file); scheduleRetry(file.getId()); log.warn(文件上传失败准备重试{}次数{}/{}, file.getId(), currentRetry 1, maxRetry); } else { // 最终失败 transactionTemplate.execute(status - { file.setUploadStatus(FAILED); file.setFailReason(重试次数耗尽 e.getMessage()); file.setUpdateTime(LocalDateTime.now()); return fileResourceService.updateById(file) 0; }); pushStatusMessage(file, FAILED); deleteTempFile(file.getTempFilePath()); } } /** * 指数退避重试调度 */ Async(fileUploadExecutor) public void scheduleRetry(Long fileId) { try { FileResource file fileResourceService.getById(fileId); int delaySeconds (int) Math.pow(2, file.getRetryCount()); Thread.sleep(delaySeconds * 1000L); asyncUpload(fileId); } catch (InterruptedException ie) { Thread.currentThread().interrupt(); log.error(重试任务被中断fileId:{}, fileId); } }六、实时状态消息推送java运行/** * 推送上传结果消息 */ private void pushStatusMessage(FileResource file, String status) { // 过滤系统内部账号 String username file.getUsername(); if (system-service.equals(username)) { return; } // 构建消息 Message message new Message(); message.setTitle(文件上传通知); message.setContent(file.getFileName() (SUCCESS.equals(status) ? 上传成功 : 上传失败)); message.setReceiver(username); message.setCreateTime(LocalDateTime.now()); message.setStatus(UNREAD); // 发布事件由 WebSocket 推送到前端 Event event new Event(); event.setType(MESSAGE); event.setData(message); eventPublisher.publishEvent(username, event); }七、实战总结线程池必须业务隔离避免某一类任务耗尽线程影响核心功能异步化是大文件上传标配可大幅提升接口吞吐量与用户体验指数退避重试能有效防止网络抖动引发的重试风暴事务控制状态保证文件记录与实际存储一致实时消息推送让用户无需轮询体验更流畅整套方案通用、轻量、无业务侵入可直接在 SpringBoot 项目中落地使用 系列导航【人生底稿 01】农村少年1995–2005【技术底稿】0137岁老码农用4台机器搭了套个人DevOps平台【产品底稿01】37 岁 Java 老码农用 Java 搭了个 AI 写作助手把自己 14 年技术文章全喂给了 AI