# DeadlineExceededError: 数据服务工作线程在 60000 毫秒后超时

- **ID:** `tensorflow/data-service-timeout`
- **领域:** tensorflow
- **类别:** network_error
- **错误码:** `DSE_TIMEOUT`
- **验证级别:** ai_generated
- **修复率:** 82%

## 根因

tf.data 服务调度器未在截止时间内收到工作线程的心跳，表明工作线程不可用或网络分区。

## 版本兼容性

| 版本 | 状态 | 引入 | 弃用 |
|------|------|------|------|
| tensorflow==2.16.1 | active | — | — |
| tensorflow-data-service==0.10.0 | active | — | — |

## 解决方案

1. ```
   Verify worker connectivity and restart the worker process with proper gRPC settings.
   ```
2. ```
   Reduce the number of parallel calls in the data pipeline to lower worker load.
   ```

## 无效尝试

- **** — Restarting the dispatcher without checking worker health does not fix the root cause of worker timeout. (80% 失败率)
- **** — Increasing the timeout hides the symptom but the underlying network issue remains. (60% 失败率)