feat: Allow configuring container logging settings on models when deploying.

PiperOrigin-RevId: 588105589

Author: vertex-sdk-bot

google/cloud/aiplatform/models.py

@@ -782,6 +782,7 @@ def deploy(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         enable_access_logging=False,
+        disable_container_logging: bool = False,
     ) -> None:
         """Deploys a Model to the Endpoint.
 
@@ -864,6 +865,9 @@ def deploy(
                 A default value of 60 will be used if not specified.
             enable_access_logging (bool):
                 Whether to enable endpoint access logging. Defaults to False.
+            disable_container_logging (bool):
+                If True, container logs from the deployed model will not be
+                written to Cloud Logging. Defaults to False.
         """
         self._sync_gca_resource_if_skipped()
 
@@ -899,6 +903,7 @@ def deploy(
             autoscaling_target_cpu_utilization=autoscaling_target_cpu_utilization,
             autoscaling_target_accelerator_duty_cycle=autoscaling_target_accelerator_duty_cycle,
             enable_access_logging=enable_access_logging,
+            disable_container_logging=disable_container_logging,
         )
 
     @base.optional_sync()
@@ -921,6 +926,7 @@ def _deploy(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         enable_access_logging=False,
+        disable_container_logging: bool = False,
     ) -> None:
         """Deploys a Model to the Endpoint.
 
@@ -997,6 +1003,9 @@ def _deploy(
                 A default value of 60 will be used if not specified.
             enable_access_logging (bool):
                 Whether to enable endpoint access logging. Defaults to False.
+            disable_container_logging (bool):
+                If True, container logs from the deployed model will not be
+                written to Cloud Logging. Defaults to False.
         """
         _LOGGER.log_action_start_against_resource(
             f"Deploying Model {model.resource_name} to", "", self
@@ -1023,6 +1032,7 @@ def _deploy(
             autoscaling_target_cpu_utilization=autoscaling_target_cpu_utilization,
             autoscaling_target_accelerator_duty_cycle=autoscaling_target_accelerator_duty_cycle,
             enable_access_logging=enable_access_logging,
+            disable_container_logging=disable_container_logging,
         )
 
         _LOGGER.log_action_completed_against_resource("model", "deployed", self)
@@ -1052,6 +1062,7 @@ def _deploy_call(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         enable_access_logging=False,
+        disable_container_logging: bool = False,
     ) -> None:
         """Helper method to deploy model to endpoint.
 
@@ -1135,6 +1146,9 @@ def _deploy_call(
                 A default value of 60 will be used if not specified.
             enable_access_logging (bool):
                 Whether to enable endpoint access logging. Defaults to False.
+            disable_container_logging (bool):
+                If True, container logs from the deployed model will not be
+                written to Cloud Logging. Defaults to False.
 
         Raises:
             ValueError: If only `accelerator_type` or `accelerator_count` is specified.
@@ -1165,6 +1179,7 @@ def _deploy_call(
             display_name=deployed_model_display_name,
             service_account=service_account,
             enable_access_logging=enable_access_logging,
+            disable_container_logging=disable_container_logging,
         )
 
         supports_automatic_resources = (
@@ -2420,6 +2435,7 @@ def deploy(
         ] = None,
         metadata: Optional[Sequence[Tuple[str, str]]] = (),
         sync=True,
+        disable_container_logging: bool = False,
     ) -> None:
         """Deploys a Model to the PrivateEndpoint.
 
@@ -2510,6 +2526,7 @@ def deploy(
             explanation_spec=explanation_spec,
             metadata=metadata,
             sync=sync,
+            disable_container_logging=disable_container_logging,
         )
 
     def undeploy(
@@ -3408,6 +3425,7 @@ def deploy(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         enable_access_logging=False,
+        disable_container_logging: bool = False,
     ) -> Union[Endpoint, PrivateEndpoint]:
         """Deploys model to endpoint. Endpoint will be created if unspecified.
 
@@ -3508,6 +3526,9 @@ def deploy(
                 A default value of 60 will be used if not specified.
             enable_access_logging (bool):
                 Whether to enable endpoint access logging. Defaults to False.
+            disable_container_logging (bool):
+                If True, container logs from the deployed model will not be
+                written to Cloud Logging. Defaults to False.
 
         Returns:
             endpoint (Union[Endpoint, PrivateEndpoint]):
@@ -3561,6 +3582,7 @@ def deploy(
             autoscaling_target_cpu_utilization=autoscaling_target_cpu_utilization,
             autoscaling_target_accelerator_duty_cycle=autoscaling_target_accelerator_duty_cycle,
             enable_access_logging=enable_access_logging,
+            disable_container_logging=disable_container_logging,
         )
 
     @base.optional_sync(return_input_arg="endpoint", bind_future_to_self=False)
@@ -3585,6 +3607,7 @@ def _deploy(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         enable_access_logging=False,
+        disable_container_logging: bool = False,
     ) -> Union[Endpoint, PrivateEndpoint]:
         """Deploys model to endpoint. Endpoint will be created if unspecified.
 
@@ -3678,6 +3701,9 @@ def _deploy(
                 A default value of 60 will be used if not specified.
             enable_access_logging (bool):
                 Whether to enable endpoint access logging. Defaults to False.
+            disable_container_logging (bool):
+                If True, container logs from the deployed model will not be
+                written to Cloud Logging. Defaults to False.
 
         Returns:
             endpoint (Union[Endpoint, PrivateEndpoint]):
@@ -3728,6 +3754,7 @@ def _deploy(
             autoscaling_target_cpu_utilization=autoscaling_target_cpu_utilization,
             autoscaling_target_accelerator_duty_cycle=autoscaling_target_accelerator_duty_cycle,
             enable_access_logging=enable_access_logging,
+            disable_container_logging=disable_container_logging,
         )
 
         _LOGGER.log_action_completed_against_resource("model", "deployed", endpoint)

google/cloud/aiplatform/preview/models.py

@@ -578,6 +578,7 @@ def deploy(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         deployment_resource_pool: Optional[DeploymentResourcePool] = None,
+        disable_container_logging: bool = False,
     ) -> None:
         """Deploys a Model to the Endpoint.
 
@@ -656,6 +657,10 @@ def deploy(
               are deployed to the same DeploymentResourcePool will be hosted in
               a shared model server. If provided, will override replica count
               arguments.
+            disable_container_logging (bool):
+              If True, container logs from the deployed model will not be
+              written to Cloud Logging. Defaults to False.
+
         """
         self._sync_gca_resource_if_skipped()
 
@@ -692,6 +697,7 @@ def deploy(
             autoscaling_target_cpu_utilization=autoscaling_target_cpu_utilization,
             autoscaling_target_accelerator_duty_cycle=autoscaling_target_accelerator_duty_cycle,
             deployment_resource_pool=deployment_resource_pool,
+            disable_container_logging=disable_container_logging,
         )
 
     @base.optional_sync()
@@ -714,6 +720,7 @@ def _deploy(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         deployment_resource_pool: Optional[DeploymentResourcePool] = None,
+        disable_container_logging: bool = False,
     ) -> None:
         """Deploys a Model to the Endpoint.
 
@@ -786,6 +793,10 @@ def _deploy(
               are deployed to the same DeploymentResourcePool will be hosted in
               a shared model server. If provided, will override replica count
               arguments.
+            disable_container_logging (bool):
+              If True, container logs from the deployed model will not be
+              written to Cloud Logging. Defaults to False.
+
         """
         _LOGGER.log_action_start_against_resource(
             f"Deploying Model {model.resource_name} to", "", self
@@ -812,6 +823,7 @@ def _deploy(
             autoscaling_target_cpu_utilization=autoscaling_target_cpu_utilization,
             autoscaling_target_accelerator_duty_cycle=autoscaling_target_accelerator_duty_cycle,
             deployment_resource_pool=deployment_resource_pool,
+            disable_container_logging=disable_container_logging,
         )
 
         _LOGGER.log_action_completed_against_resource("model", "deployed", self)
@@ -841,6 +853,7 @@ def _deploy_call(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         deployment_resource_pool: Optional[DeploymentResourcePool] = None,
+        disable_container_logging: bool = False,
     ) -> None:
         """Helper method to deploy model to endpoint.
 
@@ -920,6 +933,9 @@ def _deploy_call(
               are deployed to the same DeploymentResourcePool will be hosted in
               a shared model server. If provided, will override replica count
               arguments.
+            disable_container_logging (bool):
+              If True, container logs from the deployed model will not be
+              written to Cloud Logging. Defaults to False.
 
         Raises:
             ValueError: If only `accelerator_type` or `accelerator_count` is
@@ -953,12 +969,14 @@ def _deploy_call(
                 deploy_request_timeout=deploy_request_timeout,
                 autoscaling_target_cpu_utilization=autoscaling_target_cpu_utilization,
                 autoscaling_target_accelerator_duty_cycle=autoscaling_target_accelerator_duty_cycle,
+                disable_container_logging=disable_container_logging,
             )
 
         deployed_model = gca_endpoint_compat.DeployedModel(
             model=model.versioned_resource_name,
             display_name=deployed_model_display_name,
             service_account=service_account,
+            enable_container_logging=not disable_container_logging,
         )
 
         _LOGGER.info(model.supported_deployment_resources_types)
@@ -1237,6 +1255,7 @@ def deploy(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         deployment_resource_pool: Optional[DeploymentResourcePool] = None,
+        disable_container_logging: bool = False,
     ) -> Union[Endpoint, models.PrivateEndpoint]:
         """Deploys model to endpoint.
 
@@ -1336,6 +1355,9 @@ def deploy(
               are deployed to the same DeploymentResourcePool will be hosted in
               a shared model server. If provided, will override replica count
               arguments.
+            disable_container_logging (bool):
+              If True, container logs from the deployed model will not be
+              written to Cloud Logging. Defaults to False.
 
         Returns:
             endpoint (Union[Endpoint, models.PrivateEndpoint]):
@@ -1390,6 +1412,7 @@ def deploy(
             autoscaling_target_cpu_utilization=autoscaling_target_cpu_utilization,
             autoscaling_target_accelerator_duty_cycle=autoscaling_target_accelerator_duty_cycle,
             deployment_resource_pool=deployment_resource_pool,
+            disable_container_logging=disable_container_logging,
         )
 
     @base.optional_sync(return_input_arg="endpoint", bind_future_to_self=False)
@@ -1414,6 +1437,7 @@ def _deploy(
         autoscaling_target_cpu_utilization: Optional[int] = None,
         autoscaling_target_accelerator_duty_cycle: Optional[int] = None,
         deployment_resource_pool: Optional[DeploymentResourcePool] = None,
+        disable_container_logging: bool = False,
     ) -> Union[Endpoint, models.PrivateEndpoint]:
         """Deploys model to endpoint.
 
@@ -1505,6 +1529,9 @@ def _deploy(
               are deployed to the same DeploymentResourcePool will be hosted in
               a shared model server. If provided, will override replica count
               arguments.
+            disable_container_logging (bool):
+              If True, container logs from the deployed model will not be
+              written to Cloud Logging. Defaults to False.
 
         Returns:
             endpoint (Union[Endpoint, models.PrivateEndpoint]):
@@ -1555,6 +1582,7 @@ def _deploy(
             autoscaling_target_cpu_utilization=autoscaling_target_cpu_utilization,
             autoscaling_target_accelerator_duty_cycle=autoscaling_target_accelerator_duty_cycle,
             deployment_resource_pool=deployment_resource_pool,
+            disable_container_logging=disable_container_logging,
         )
 
         _LOGGER.log_action_completed_against_resource("model", "deployed", endpoint)

tests/unit/aiplatform/test_endpoints.py

@@ -1676,6 +1676,42 @@ def test_deploy_with_max_replica_count(self, deploy_model_mock, sync):
             timeout=None,
         )
 
+    @pytest.mark.usefixtures("get_endpoint_mock", "get_model_mock")
+    @pytest.mark.parametrize("sync", [True, False])
+    def test_deploy_disable_container_logging(self, deploy_model_mock, sync):
+        test_endpoint = models.Endpoint(_TEST_ENDPOINT_NAME)
+        test_model = models.Model(_TEST_ID)
+        test_model._gca_resource.supported_deployment_resources_types.append(
+            aiplatform.gapic.Model.DeploymentResourcesType.AUTOMATIC_RESOURCES
+        )
+        test_endpoint.deploy(
+            test_model,
+            sync=sync,
+            deploy_request_timeout=None,
+            disable_container_logging=True,
+        )
+
+        if not sync:
+            test_endpoint.wait()
+
+        automatic_resources = gca_machine_resources.AutomaticResources(
+            min_replica_count=1,
+            max_replica_count=1,
+        )
+        deployed_model = gca_endpoint.DeployedModel(
+            automatic_resources=automatic_resources,
+            model=test_model.resource_name,
+            display_name=None,
+            disable_container_logging=True,
+        )
+        deploy_model_mock.assert_called_once_with(
+            endpoint=test_endpoint.resource_name,
+            deployed_model=deployed_model,
+            traffic_split={"0": 100},
+            metadata=(),
+            timeout=None,
+        )
+
     @pytest.mark.usefixtures("get_endpoint_mock", "get_model_mock", "get_drp_mock")
     @pytest.mark.parametrize("sync", [True, False])
     def test_preview_deploy_with_deployment_resource_pool(
@@ -1701,6 +1737,7 @@ def test_preview_deploy_with_deployment_resource_pool(
             shared_resources=_TEST_DRP_NAME,
             model=test_model.resource_name,
             display_name=None,
+            enable_container_logging=True,
         )
         preview_deploy_model_mock.assert_called_once_with(
             endpoint=test_endpoint.resource_name,

tests/unit/aiplatform/test_models.py

@@ -2171,6 +2171,47 @@ def test_deploy_no_endpoint_with_network(self, deploy_model_mock):
             timeout=None,
         )
 
+    @pytest.mark.usefixtures("get_endpoint_mock", "get_model_mock")
+    @pytest.mark.parametrize("sync", [True, False])
+    def test_deploy_disable_container_logging(self, deploy_model_mock, sync):
+
+        test_model = models.Model(_TEST_ID)
+        test_model._gca_resource.supported_deployment_resources_types.append(
+            aiplatform.gapic.Model.DeploymentResourcesType.AUTOMATIC_RESOURCES
+        )
+
+        test_endpoint = models.Endpoint(_TEST_ID)
+
+        assert (
+            test_model.deploy(
+                test_endpoint,
+                disable_container_logging=True,
+                sync=sync,
+            )
+            == test_endpoint
+        )
+
+        if not sync:
+            test_endpoint.wait()
+
+        automatic_resources = gca_machine_resources.AutomaticResources(
+            min_replica_count=1,
+            max_replica_count=1,
+        )
+        deployed_model = gca_endpoint.DeployedModel(
+            automatic_resources=automatic_resources,
+            model=test_model.resource_name,
+            display_name=None,
+            disable_container_logging=True,
+        )
+        deploy_model_mock.assert_called_once_with(
+            endpoint=test_endpoint.resource_name,
+            deployed_model=deployed_model,
+            traffic_split={"0": 100},
+            metadata=(),
+            timeout=None,
+        )
+
     @pytest.mark.usefixtures(
         "get_model_mock", "get_drp_mock", "create_endpoint_mock", "get_endpoint_mock"
     )
@@ -2196,6 +2237,7 @@ def test_preview_deploy_with_deployment_resource_pool(
             shared_resources=_TEST_DRP_NAME,
             model=test_model.resource_name,
             display_name=None,
+            enable_container_logging=True,
         )
         preview_deploy_model_mock.assert_called_once_with(
             endpoint=test_endpoint.resource_name,