大模型API错误处理：构建防弹级AI应用

class LLMAPIClient {
  constructor(config) {
    this.config = config;
    this.retryConfig = {
      maxRetries: 3,
      retryableErrors: [429, 500, 502, 503, 504],
      backoffMultiplier: 2,
      maxBackoff: 60000
    };
  }

  async callAPI(params) {
    let lastError;
    
    for (let attempt = 0; attempt <= this.retryConfig.maxRetries; attempt++) {
      try {
        // 添加超时控制
        const response = await this.makeRequest(params, {
          timeout: 30000,
          signal: AbortSignal.timeout(30000)
        });
        
        // 验证响应
        this.validateResponse(response);
        
        // 记录成功
        this.metrics.recordSuccess(attempt);
        
        return response;
        
      } catch (error) {
        lastError = error;
        
        // 记录错误
        this.logError(error, attempt);
        
        // 判断是否可重试
        if (!this.shouldRetry(error, attempt)) {
          throw this.wrapError(error);
        }
        
        // 等待后重试
        await this.waitBeforeRetry(error, attempt);
        
        // 错误恢复策略
        params = await this.applyRecoveryStrategy(error, params);
      }
    }
    
    throw new MaxRetriesError(lastError);
  }

  shouldRetry(error, attempt) {
    // 不可重试的错误
    if (error.code === 'invalid_api_key' || 
        error.code === 'insufficient_quota') {
      return false;
    }
    
    // 达到最大重试次数
    if (attempt >= this.retryConfig.maxRetries) {
      return false;
    }
    
    // 检查HTTP状态码
    return this.retryConfig.retryableErrors.includes(error.status);
  }

  async waitBeforeRetry(error, attempt) {
    let delay;
    
    // 优先使用服务器指定的重试时间
    if (error.retryAfter) {
      delay = error.retryAfter * 1000;
    } else {
      // 指数退避 + 抖动
      delay = Math.min(
        1000 * Math.pow(this.retryConfig.backoffMultiplier, attempt),
        this.retryConfig.maxBackoff
      );
      delay += Math.random() * 1000; // 添加随机抖动
    }
    
    await new Promise(resolve => setTimeout(resolve, delay));
  }

  async applyRecoveryStrategy(error, params) {
    switch (error.code) {
      case 'context_length_exceeded':
        // 压缩上下文
        return {
          ...params,
          messages: this.compressMessages(params.messages)
        };
        
      case 'model_overloaded':
        // 降级到更快的模型
        return {
          ...params,
          model: this.getFallbackModel(params.model)
        };
        
      default:
        return params;
    }
  }
}