修复首次会话等待与上游饱和重试

tests/unit/ai-proxy-routes.test.ts

@@ -336,20 +336,21 @@ describe('ai proxy routes', () => {
     expect(response.body()).toContain('works_square_gateway_authorize_failed');
   });
 
-  it('logs a sanitized one-api error summary when the upstream group is saturated', async () => {
+  it('maps explicit upstream group saturation to a non-retryable response status', async () => {
     seedWorksSquareAIGatewayCredential({
       accessToken: 'ws-ai-token',
       expiresIn: 3600,
       oneApiBaseUrl: 'https://one-api.example.com/v1',
     });
+    const body = JSON.stringify({
+      error: {
+        message: '当前分组上游负载已饱和，请稍后再试 (request id: req-saturated)',
+        code: 'rate_limit_exceeded',
+        type: 'one_api_error',
+      },
+    });
     const fetchMock = vi.fn().mockResolvedValueOnce(
-      new Response(JSON.stringify({
-        error: {
-          message: '当前分组上游负载已饱和，请稍后再试 (request id: req-saturated)',
-          code: 'rate_limit_exceeded',
-          type: 'one_api_error',
-        },
-      }), {
+      new Response(body, {
         status: 429,
         headers: { 'content-type': 'application/json' },
       }),
@@ -364,8 +365,10 @@ describe('ai proxy routes', () => {
       {} as never,
     );
 
-    expect(response.statusCode).toBe(429);
-    expect(response.body()).toContain('当前分组上游负载已饱和');
+    expect(fetchMock).toHaveBeenCalledOnce();
+    expect(response.statusCode).toBe(400);
+    expect(response.header('content-type')).toContain('application/json');
+    expect(response.body()).toBe(body);
     expect(loggerWarnMock).toHaveBeenCalledWith(
       '[ai-proxy] One-api returned non-success response',
       expect.objectContaining({
@@ -380,6 +383,40 @@ describe('ai proxy routes', () => {
     expect(JSON.stringify(loggerWarnMock.mock.calls)).not.toContain('ws-ai-token');
   });
 
+  it('preserves a generic rate-limit response without explicit saturation evidence', async () => {
+    seedWorksSquareAIGatewayCredential({
+      accessToken: 'ws-ai-token',
+      expiresIn: 3600,
+      oneApiBaseUrl: 'https://one-api.example.com/v1',
+    });
+    const body = JSON.stringify({
+      error: {
+        message: 'Rate limit exceeded',
+        code: 'rate_limit_exceeded',
+        type: 'one_api_error',
+      },
+    });
+    const fetchMock = vi.fn().mockResolvedValueOnce(
+      new Response(body, {
+        status: 429,
+        headers: { 'content-type': 'application/json' },
+      }),
+    );
+    vi.stubGlobal('fetch', fetchMock);
+    const response = createResponse();
+
+    await handleAiProxyRoutes(
+      createRequest('POST', { model: 'qwen3.7-max', messages: [] }),
+      response.res,
+      new URL('http://127.0.0.1:13210/api/ai-proxy/v1/chat/completions'),
+      {} as never,
+    );
+
+    expect(fetchMock).toHaveBeenCalledOnce();
+    expect(response.statusCode).toBe(429);
+    expect(response.body()).toBe(body);
+  });
+
   it('strips decoded compression headers from proxied responses', async () => {
     seedWorksSquareAIGatewayCredential({
       accessToken: 'ws-ai-token',