fix(tests): add OpenAI client connection cleanup to prevent CI hangs (#4119)

# What does this PR do?

Add explicit connection cleanup and shorter timeouts to OpenAI client
fixtures. Fixes CI deadlock after 25+ tests due to connection pool
exhaustion. Also adds 60s timeout to test_conversation_context_loading
as safety net.

## Test Plan

tests pass

Signed-off-by: Charlie Doern <[email protected]>

Author: cdoern

tests/integration/fixtures/common.py

@@ -323,7 +323,13 @@ def require_server(llama_stack_client):
 @pytest.fixture(scope="session")
 def openai_client(llama_stack_client, require_server):
     base_url = f"{llama_stack_client.base_url}/v1"
-    return OpenAI(base_url=base_url, api_key="fake")
+    client = OpenAI(base_url=base_url, api_key="fake", max_retries=0, timeout=30.0)
+    yield client
+    # Cleanup: close HTTP connections
+    try:
+        client.close()
+    except Exception:
+        pass
 
 
 @pytest.fixture(params=["openai_client", "client_with_models"])

tests/integration/responses/fixtures/fixtures.py

@@ -115,7 +115,15 @@ def openai_client(base_url, api_key, provider):
         client = LlamaStackAsLibraryClient(config, skip_logger_removal=True)
         return client
 
-    return OpenAI(
+    client = OpenAI(
         base_url=base_url,
         api_key=api_key,
+        max_retries=0,
+        timeout=30.0,
     )
+    yield client
+    # Cleanup: close HTTP connections
+    try:
+        client.close()
+    except Exception:
+        pass

tests/integration/responses/test_conversation_responses.py

@@ -65,8 +65,14 @@ def test_conversation_multi_turn_and_streaming(self, openai_client, text_model_i
         conversation_items = openai_client.conversations.items.list(conversation.id)
         assert len(conversation_items.data) >= 4  # 2 user + 2 assistant messages
 
+    @pytest.mark.timeout(60, method="thread")
     def test_conversation_context_loading(self, openai_client, text_model_id):
-        """Test that conversation context is properly loaded for responses."""
+        """Test that conversation context is properly loaded for responses.
+
+        Note: 60s timeout added due to CI-specific deadlock in pytest/OpenAI client/httpx
+        after running 25+ tests. Hangs before first HTTP request is made. Works fine locally.
+        Investigation needed: connection pool exhaustion or event loop state issue.
+        """
         conversation = openai_client.conversations.create(
             items=[
                 {"type": "message", "role": "user", "content": "My name is Alice. I like to eat apples."},