aws / sagemaker-python-sdk

A library for training and deploying machine learning models on Amazon SageMaker
https://sagemaker.readthedocs.io/
Apache License 2.0
2.11k stars 1.14k forks source link

Explicit parameter to skip endpoint validation when creating a monitoring schedule #4873

Open luke-gerschwitz opened 2 months ago

luke-gerschwitz commented 2 months ago

Describe the feature you'd like An explicit parameter to skip endpoint validation when calling create_monitoring_schedule and passing an EndpointInput object. Currently SageMaker verifies that an endpoint exists in the same workspace as the monitoring schedule, however this causes issues when creating a monitoring schedule to monitor an endpoint that exists in a different workspace.

How would this feature be used? Please describe. When creating a cross-account monitoring schedule, the user should be able to explicitly skip the endpoint validation that occurs (Or SageMaker should validate the existence of cross-account endpoints that the execution role has access to).

Describe alternatives you've considered Currently the only workaround is to create a dummy endpoint in the same workspace that the monitoring schedule is being created in. This endpoint then needs to have an endpoint configuration attached that has DataCaptureConfig.DestinationS3Uri value set to the data capture bucket in the workspace containing the actual endpoint. This is cumbersome and inefficient at scale.

Additional context Add any other context or screenshots about the feature request here.