feat: expose gcf max timeout in remote_function

bigframes/functions/remote_function.py

@@ -341,7 +341,9 @@ def generate_cloud_function_code(self, def_, dir, package_requirements=None):
         entry_point = self.generate_cloud_function_main_code(def_, dir)
         return entry_point
 
-    def create_cloud_function(self, def_, cf_name, package_requirements=None):
+    def create_cloud_function(
+        self, def_, cf_name, package_requirements=None, cloud_function_timeout=600
+    ):
         """Create a cloud function from the given user defined function."""
 
         # Build and deploy folder structure containing cloud function
@@ -409,7 +411,14 @@ def create_cloud_function(self, def_, cf_name, package_requirements=None):
             )
             function.service_config = functions_v2.ServiceConfig()
             function.service_config.available_memory = "1024M"
-            function.service_config.timeout_seconds = 600
+            if cloud_function_timeout is not None:
+                if cloud_function_timeout > 1200:
+                    raise ValueError(
+                        "BigQuery remote function can wait only up to 20 minutes"
+                        ", see for more details "
+                        "https://cloud.google.com/bigquery/quotas#remote_function_limits."
+                    )
+                function.service_config.timeout_seconds = cloud_function_timeout
             function.service_config.service_account_email = (
                 self._cloud_function_service_account
             )
@@ -456,6 +465,7 @@ def provision_bq_remote_function(
         name,
         package_requirements,
         max_batching_rows,
+        cloud_function_timeout,
     ):
         """Provision a BigQuery remote function."""
         # If reuse of any existing function with the same name (indicated by the
@@ -477,7 +487,7 @@ def provision_bq_remote_function(
         # Create the cloud function if it does not exist
         if not cf_endpoint:
             cf_endpoint = self.create_cloud_function(
-                def_, cloud_function_name, package_requirements
+                def_, cloud_function_name, package_requirements, cloud_function_timeout
             )
         else:
             logger.info(f"Cloud function {cloud_function_name} already exists.")
@@ -631,6 +641,7 @@ def remote_function(
     cloud_function_kms_key_name: Optional[str] = None,
     cloud_function_docker_repository: Optional[str] = None,
     max_batching_rows: Optional[int] = 1000,
+    cloud_function_timeout: Optional[int] = 600,
 ):
     """Decorator to turn a user defined function into a BigQuery remote function.
 
@@ -756,6 +767,16 @@ def remote_function(
             `None` can be passed to let BQ remote functions service apply
             default batching. See for more details
             https://cloud.google.com/bigquery/docs/remote-functions#limiting_number_of_rows_in_a_batch_request.
+        cloud_function_timeout (int, Optional):
+            The maximum amount of time (in seconds) BigQuery should wait for
+            the cloud function to return a response. See for more details
+            https://cloud.google.com/functions/docs/configuring/timeout.
+            Please note that even though the cloud function (2nd gen) itself
+            allows seeting up to 60 minutes of timeout, BigQuery remote
+            function can wait only up to 20 minutes, see for more details
+            https://cloud.google.com/bigquery/quotas#remote_function_limits.
+            By default BigQuery DataFrames uses a 10 minute timeout. `None`
+            can be passed to let the cloud functions default timeout take effect.
     """
     import bigframes.pandas as bpd
 
@@ -880,6 +901,7 @@ def wrapper(f):
             name,
             packages,
             max_batching_rows,
+            cloud_function_timeout,
         )
 
         # TODO: Move ibis logic to compiler step

bigframes/pandas/__init__.py

@@ -644,6 +644,7 @@ def remote_function(
     cloud_function_kms_key_name: Optional[str] = None,
     cloud_function_docker_repository: Optional[str] = None,
     max_batching_rows: Optional[int] = 1000,
+    cloud_function_timeout: Optional[int] = 600,
 ):
     return global_session.with_default_session(
         bigframes.session.Session.remote_function,
@@ -658,6 +659,7 @@ def remote_function(
         cloud_function_kms_key_name=cloud_function_kms_key_name,
         cloud_function_docker_repository=cloud_function_docker_repository,
         max_batching_rows=max_batching_rows,
+        cloud_function_timeout=cloud_function_timeout,
     )
 
 

bigframes/session/__init__.py

@@ -1545,6 +1545,7 @@ def remote_function(
         cloud_function_kms_key_name: Optional[str] = None,
         cloud_function_docker_repository: Optional[str] = None,
         max_batching_rows: Optional[int] = 1000,
+        cloud_function_timeout: Optional[int] = 600,
     ):
         """Decorator to turn a user defined function into a BigQuery remote function. Check out
         the code samples at: https://cloud.google.com/bigquery/docs/remote-functions#bigquery-dataframes.
@@ -1648,6 +1649,16 @@ def remote_function(
                 `None` can be passed to let BQ remote functions service apply
                 default batching. See for more details
                 https://cloud.google.com/bigquery/docs/remote-functions#limiting_number_of_rows_in_a_batch_request.
+            cloud_function_timeout (int, Optional):
+                The maximum amount of time (in seconds) BigQuery should wait for
+                the cloud function to return a response. See for more details
+                https://cloud.google.com/functions/docs/configuring/timeout.
+                Please note that even though the cloud function (2nd gen) itself
+                allows seeting up to 60 minutes of timeout, BigQuery remote
+                function can wait only up to 20 minutes, see for more details
+                https://cloud.google.com/bigquery/quotas#remote_function_limits.
+                By default BigQuery DataFrames uses a 10 minute timeout. `None`
+                can be passed to let the cloud functions default timeout take effect.
         Returns:
             callable: A remote function object pointing to the cloud assets created
             in the background to support the remote execution. The cloud assets can be
@@ -1670,6 +1681,7 @@ def remote_function(
             cloud_function_kms_key_name=cloud_function_kms_key_name,
             cloud_function_docker_repository=cloud_function_docker_repository,
             max_batching_rows=max_batching_rows,
+            cloud_function_timeout=cloud_function_timeout,
         )
 
     def read_gbq_function(

tests/system/large/test_remote_function.py

@@ -1336,3 +1336,52 @@ def square(x):
         cleanup_remote_function_assets(
             session.bqclient, session.cloudfunctionsclient, square_remote
         )
+
+
+@pytest.mark.parametrize(
+    ("timeout_args", "effective_gcf_timeout"),
+    [
+        pytest.param({}, 600, id="no-set"),
+        pytest.param({"cloud_function_timeout": None}, 60, id="set-None"),
+        pytest.param({"cloud_function_timeout": 1200}, 1200, id="set-max-allowed"),
+    ],
+)
+@pytest.mark.flaky(retries=2, delay=120)
+def test_remote_function_gcf_timeout(
+    session, scalars_dfs, timeout_args, effective_gcf_timeout
+):
+    try:
+
+        def square(x):
+            return x * x
+
+        square_remote = session.remote_function(
+            [int], int, reuse=False, **timeout_args
+        )(square)
+
+        # Assert that the GCF is created with the intended maximum timeout
+        gcf = session.cloudfunctionsclient.get_function(
+            name=square_remote.bigframes_cloud_function
+        )
+        assert gcf.service_config.timeout_seconds == effective_gcf_timeout
+
+        scalars_df, scalars_pandas_df = scalars_dfs
+
+        bf_result = scalars_df["int64_too"].apply(square_remote).to_pandas()
+        pd_result = scalars_pandas_df["int64_too"].apply(square)
+
+        pandas.testing.assert_series_equal(bf_result, pd_result, check_dtype=False)
+    finally:
+        # clean up the gcp assets created for the remote function
+        cleanup_remote_function_assets(
+            session.bqclient, session.cloudfunctionsclient, square_remote
+        )
+
+
+@pytest.mark.flaky(retries=2, delay=120)
+def test_remote_function_gcf_timeout_max_supported_exceeded(session):
+    with pytest.raises(ValueError):
+
+        @session.remote_function([int], int, reuse=False, cloud_function_timeout=1201)
+        def square(x):
+            return x * x