fix(gooddata-pandas): drain HTTP body to enable keep-alive reuse

risk: low

Author: Martozar

packages/gooddata-sdk/src/gooddata_sdk/compute/model/execution.py

@@ -1,6 +1,7 @@
 # (C) 2022 GoodData Corporation
 from __future__ import annotations
 
+import io
 import logging
 from typing import TYPE_CHECKING, Any, Union
 
@@ -408,9 +409,14 @@ def read_result_arrow(self) -> pyarrow.Table:
             _return_http_data_only=True,
         )
         try:
-            return _ipc.open_stream(response).read_all()
+            # Read the full HTTP response body before releasing the connection.
+            # pyarrow's IPC stream reader stops at the Arrow EOS marker and may leave
+            # the HTTP chunked-encoding terminator unread, which puts the connection in
+            # Request-sent state and breaks HTTP keep-alive reuse.
+            data = response.read()
         finally:
             response.release_conn()
+        return _ipc.open_stream(io.BytesIO(data)).read_all()
 
     def cancel(self) -> None:
         """