Skip to content

Support for SageMaker Notebook Operators #32702

Description

@shubham22

Description

Today, Amazon provider package supports SageMaker operators for a few operations, like training, tuning, pipelines, but it lacks the support for SageMaker Notebook instances. Boto3 provides necessary APIs to create notebook instance, start notebook instance, stop network instance and delete notebook instance. Leveraging these APIs, we should add new operators to SageMaker set under Amazon provider. At the same time, having a sensor (synchronous as well as deferrable) for notebook instance execution that utilizes describe notebook instance and waits for Stopped/Failed status would help with observability of the execution.

Use case/motivation

Data Scientists are orchestrating ML use cases via Apache Airflow. A key component of ML use cases is running Jupyter Notebook on SageMaker. Having built-in operators and sensors would make it easy for Airflow users to run Notebook instances on SageMaker.

Related issues

No response

Are you willing to submit a PR?

  • Yes I am willing to submit a PR!

Code of Conduct

Metadata

Metadata

Assignees

Type

No type

Projects

No projects

Milestone

No milestone

Relationships

None yet

Development

No branches or pull requests

Issue actions