fix: update remaining e2e test timeouts from 10s to 30s

Updated timeout values in 5 test files to prevent timeout failures:
- 00-client-request-test.py
- 04-cache-test.py
- 06-tools-test.py
- 07-model-selection-test.py
- 09-error-handling-test.py

This should resolve the remaining timeout issues seen in local testing.

Author: yossiovadia

e2e-tests/00-client-request-test.py

@@ -176,14 +176,14 @@ def test_malformed_request(self):
 
         try:
             response = self._make_request(
-                payload, timeout=10
+                payload, timeout=30
             )  # Reduced timeout for error cases
             if not response:
                 response = requests.post(
                     f"{ENVOY_URL}{OPENAI_ENDPOINT}",
                     headers={"Content-Type": "application/json"},
                     json=payload,
-                    timeout=10,
+                    timeout=30,
                 )
 
             passed = 400 <= response.status_code < 500

e2e-tests/04-cache-test.py

@@ -136,7 +136,7 @@ def test_cache_hit_with_identical_query(self):
         # First request
         self.print_subtest_header("First Request (Expected Cache Miss)")
         response1 = requests.post(
-            f"{ENVOY_URL}{OPENAI_ENDPOINT}", headers=headers, json=payload, timeout=10
+            f"{ENVOY_URL}{OPENAI_ENDPOINT}", headers=headers, json=payload, timeout=30
         )
 
         response1_json = response1.json()
@@ -157,7 +157,7 @@ def test_cache_hit_with_identical_query(self):
         # Second identical request
         self.print_subtest_header("Second Request (Expected Cache Hit)")
         response2 = requests.post(
-            f"{ENVOY_URL}{OPENAI_ENDPOINT}", headers=headers, json=payload, timeout=10
+            f"{ENVOY_URL}{OPENAI_ENDPOINT}", headers=headers, json=payload, timeout=30
         )
 
         response2_json = response2.json()
@@ -237,7 +237,7 @@ def test_cache_hit_with_similar_query(self):
         headers = {"Content-Type": "application/json", "X-Session-ID": session_id}
 
         response1 = requests.post(
-            f"{ENVOY_URL}{OPENAI_ENDPOINT}", headers=headers, json=payload1, timeout=10
+            f"{ENVOY_URL}{OPENAI_ENDPOINT}", headers=headers, json=payload1, timeout=30
         )
 
         response1_json = response1.json()
@@ -273,7 +273,7 @@ def test_cache_hit_with_similar_query(self):
         )
 
         response2 = requests.post(
-            f"{ENVOY_URL}{OPENAI_ENDPOINT}", headers=headers, json=payload2, timeout=10
+            f"{ENVOY_URL}{OPENAI_ENDPOINT}", headers=headers, json=payload2, timeout=30
         )
 
         response2_json = response2.json()

e2e-tests/06-tools-test.py

@@ -183,7 +183,7 @@ def test_specific_tool_selection(self):
                     f"{ENVOY_URL}{OPENAI_ENDPOINT}",
                     headers=headers,
                     json=payload,
-                    timeout=10,
+                    timeout=30,
                 )
 
                 # Tool selection should work regardless of vLLM backend availability
@@ -259,7 +259,7 @@ def test_no_tool_requests(self):
                     f"{ENVOY_URL}{OPENAI_ENDPOINT}",
                     headers=headers,
                     json=payload,
-                    timeout=10,
+                    timeout=30,
                 )
 
                 passed = response.status_code in [200, 503]
@@ -331,7 +331,7 @@ def test_tools_configuration_validation(self):
             f"{ENVOY_URL}{OPENAI_ENDPOINT}",
             headers=headers,
             json=payload,
-            timeout=10,
+            timeout=30,
         )
 
         passed = response.status_code in [200, 503]
@@ -452,7 +452,7 @@ def test_tools_selection_consistency(self):
                 f"{ENVOY_URL}{OPENAI_ENDPOINT}",
                 headers=headers,
                 json=payload,
-                timeout=10,
+                timeout=30,
             )
 
             # Record the response status for consistency checking

e2e-tests/07-model-selection-test.py

@@ -214,7 +214,7 @@ def test_category_based_model_selection(self):
                     f"{ENVOY_URL}{OPENAI_ENDPOINT}",
                     headers=headers,
                     json=payload,
-                    timeout=10,
+                    timeout=30,
                 )
 
                 passed = response.status_code in [200, 503]
@@ -293,7 +293,7 @@ def test_reasoning_mode_selection(self):
                     f"{ENVOY_URL}{OPENAI_ENDPOINT}",
                     headers=headers,
                     json=payload,
-                    timeout=10,
+                    timeout=30,
                 )
 
                 passed = response.status_code in [200, 503]
@@ -365,7 +365,7 @@ def test_model_fallback_behavior(self):
             f"{ENVOY_URL}{OPENAI_ENDPOINT}",
             headers=headers,
             json=payload,
-            timeout=10,
+            timeout=30,
         )
 
         # Fallback should work, though may get 503 if no vLLM backend
@@ -490,7 +490,7 @@ def test_model_selection_consistency(self):
                 f"{ENVOY_URL}{OPENAI_ENDPOINT}",
                 headers=headers,
                 json=payload,
-                timeout=10,
+                timeout=30,
             )
 
             # Record the response status and model for consistency checking

e2e-tests/09-error-handling-test.py

@@ -244,7 +244,7 @@ def test_malformed_requests(self):
                     f"{ENVOY_URL}{OPENAI_ENDPOINT}",
                     headers=headers,
                     json=test_case["payload"],
-                    timeout=10,
+                    timeout=30,
                 )
 
                 min_status, max_status = test_case["expected_status_range"]
@@ -467,7 +467,7 @@ def test_invalid_content_types(self):
                     f"{ENVOY_URL}{OPENAI_ENDPOINT}",
                     headers=headers,
                     data=test_case["body"],
-                    timeout=10,
+                    timeout=30,
                 )
 
                 # Invalid content types should typically return 4xx errors
@@ -517,7 +517,7 @@ def test_error_response_format(self):
             f"{ENVOY_URL}{OPENAI_ENDPOINT}",
             headers=headers,
             json=payload,
-            timeout=10,
+            timeout=30,
         )
 
         passed = response.status_code >= 400