错误信息: upstream error: do request failed (request id: 20260103204432591660753Mp1r...

2026年01月03日 20:50 processing

错误信息

错误名称: AI_APICallError 错误信息: upstream error: do request failed (request id: 20260103204432591660753Mp1rvxbJ) 堆栈信息: AI_APICallError: upstream error: do request failed (request id: 20260103204432591660753Mp1rvxbJ) at file:///C:/Users/86153/AppData/Local/Programs/Cherry%20Studio/resources/app.asar/out/renderer/assets/dist-K3A05YNJ.js:543:11 at async postToApi (file:///C:/Users/86153/AppData/Local/Programs/Cherry%20Studio/resources/app.asar/out/renderer/assets/dist-K3A05YNJ.js:439:24) at async OpenAICompatibleChatLanguageModel$1.doStream (file:///C:/Users/86153/AppData/Local/Programs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-e4Pmzkul.js:253086:48) at async fn (file:///C:/Users/86153/AppData/Local/Programs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-e4Pmzkul.js:37120:17) at async file:///C:/Users/86153/AppData/Local/Programs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-e4Pmzkul.js:34498:19 at async _retryWithExponentialBackoff (file:///C:/Users/86153/AppData/Local/Programs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-e4Pmzkul.js:34600:10) at async streamStep (file:///C:/Users/86153/AppData/Local/Programs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-e4Pmzkul.js:37088:108) at async fn (file:///C:/Users/86153/AppData/Local/Programs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-e4Pmzkul.js:37367:5) at async file:///C:/Users/86153/AppData/Local/Programs/Cherry%20Studio/resources/app.asar/out/renderer/assets/store-e4Pmzkul.js:34498:19 错误原因: "[undefined]" 状态码: 500 请求路径: https://www.dmxapi.cn/v1/chat/completions 请求体: { "model": "claude-3-5-sonnet-20241022", "user": "[undefined]", "max_tokens": "[undefined]", "temperature": 1, "top_p": "[undefined]", "frequency_penalty": "[undefined]", "presence_penalty": "[undefined]", "response_format": "[undefined]", "stop": "[undefined]", "seed": "[undefined]", "serviceTier": "[undefined]", "reasoning_effort": "[undefined]", "verbosity": "[undefined]", "messages": [ { "role": "user", "content": "在宏观的建筑生长角度，从控制建筑的生长密度和控制建筑的生长类型角度出发，分别有一些什么样的约束条件，（比如物理环境的舒适度需求，人体的舒适度需求），请分别给我至少约束条件" } ], "tools": "[undefined]", "tool_choice": "[undefined]", "stream": true, "stream_options": { "include_usage": true } } 响应首部: { "cache-control": "no-cache", "content-type": "text/event-stream", "date": "Sat, 03 Jan 2026 12:49:46 GMT", "document-policy": "include-js-call-stacks-in-crash-reports", "server": "nginx", "x-rixapi-request-id": "20260103204938359922572uhfpm1Ae" } 响应内容: { "error": { "message": "upstream error: do request failed (request id: 20260103204432591660753Mp1rvxbJ)", "type": "rix_api_error", "param": "", "code": null } } 数据: { "error": { "message": "upstream error: do request failed (request id: 20260103204432591660753Mp1rvxbJ)", "type": "rix_api_error", "param": "", "code": null } }

AI 解决方案

AI API调用错误分析报告

1. 错误翻译

原文：upstream error: do request failed
译文：上游错误：执行请求失败

2. 原因分析

这是一个典型的上游API服务错误，核心问题如下：

主要问题：

上游服务故障：DMXAPI平台的后端服务在处理请求时失败

状态码500：服务器内部错误，表明问题在服务端而非客户端

请求参数问题：多个参数值为"[undefined]"字符串而非实际的undefined或有效值

关键发现：

请求体格式异常：多个参数被错误地设置为字符串"[undefined]"而不是真正的undefined

代理层错误：Cherry Studio通过DMXAPI调用Claude-3-5-Sonnet模型时，中间层出现问题

请求ID完整链路：20260103204432591660753Mp1rvxbJ → 20260103204938359922572uhfpm1Ae

3. 解决方案

客户端修复（Cherry Studio端）：

// 修复前的错误参数设置
const wrongParams = {
  user: "[undefined]",          // ❌ 错误：字符串"[undefined]"
  max_tokens: "[undefined]",    // ❌ 应该为undefined或数字
  // ... 其他参数
};

// 修复后的正确设置
const correctParams = {
  user: undefined,              // ✅ 正确：真正的undefined
  max_tokens: undefined,        // ✅ 或设置为有效数值如4096
  temperature: 1,
  top_p: undefined,             // ✅ 使用undefined而不是字符串
  frequency_penalty: undefined,
  presence_penalty: undefined,
  response_format: undefined,
  stop: undefined,
  seed: undefined,
  serviceTier: undefined,
  reasoning_effort: undefined,
  verbosity: undefined,
  model: "claude-3-5-sonnet-20241022",
  messages: [
    {
      role: "user",
      content: "在宏观的建筑生长角度，从控制建筑的生长密度和控制建筑的生长类型角度出发，分别有一些什么样的约束条件，（比如物理环境的舒适度需求，人体的舒适度需求），请分别给我至少约束条件"
    }
  ],
  tools: undefined,
  tool_choice: undefined,
  stream: true,
  stream_options: {
    include_usage: true
  }
};

临时处理方案：

// 1. 清理请求参数
function cleanApiParams(params) {
  const cleaned = { ...params };
  
  // 移除所有值为"[undefined]"的参数
  Object.keys(cleaned).forEach(key => {
    if (cleaned[key] === "[undefined]") {
      delete cleaned[key];
    }
  });
  
  return cleaned;
}

// 2. 重试机制
async function callApiWithRetry(apiCall, maxRetries = 3) {
  for (let i = 0; i < maxRetries; i++) {
    try {
      return await apiCall();
    } catch (error) {
      if (error.statusCode === 500 && i < maxRetries - 1) {
        console.log(`API调用失败，${i+1}秒后重试...`);
        await new Promise(resolve => setTimeout(resolve, 1000 * (i + 1)));
      } else {
        throw error;
      }
    }
  }
}

4. ️ 预防措施

️ 最佳实践：

参数验证：在发送API请求前，验证所有参数类型

   function validateApiParams(params) {
     const allowedUndefined = ['user', 'tools', 'tool_choice'];
     
     Object.entries(params).forEach(([key, value]) => {
       if (value === "[undefined]" || (value === undefined && !allowedUndefined.includes(key))) {
         console.warn(`参数 ${key} 存在潜在问题:`, value);
       }
     });
   }

错误处理优化：添加更详细的错误日志记录

   try {
     const response = await fetch(apiUrl, options);
     if (!response.ok) {
       const errorText = await response.text();
       console.error(`API响应异常:
         状态码: ${response.status}
         请求ID: ${response.headers.get('x-request-id')}
         错误详情: ${errorText}`);
     }
     return await response.json();
   } catch (error) {
     console.error('API调用失败:', {
       错误类型: error.name,
       错误信息: error.message,
       请求URL: apiUrl,
       时间戳: new Date().toISOString()
     });
     throw error;
   }

监控建议：

记录所有请求ID：便于跟踪完整的请求链路

设置API健康检查：定期测试DMXAPI服务可用性

使用参数清理中间件：确保发送到API的数据格式正确

调试工具推荐：

Postman/Insomnia：手动测试API端点

Charles Proxy/Fiddler：抓包分析请求响应

Chrome DevTools Network面板：实时监控API调用

关键提示：此错误主要由服务端引起，但客户端不正确的参数格式可能触发服务端的异常处理路径。建议先修复参数格式，如果问题持续存在，联系DMXAPI技术支持并提供完整的请求ID链路。

分析新问题查看更多