Tags not propogating to Endpoints and Models

[AnshumanRanjan]

Please fill out the form below.

System Information

• Framework (e.g. TensorFlow) / Algorithm (e.g. KMeans): Sklearn
• Framework Version:NA
• Python Version: 3.6.6
• CPU or GPU: CPU
• Python SDK Version:1.16.3
• Are you using a custom image: NO

Describe the problem

Based on my testing , tags parameter defined in the Estimator objects are not propagated to Models and Endpoints . Although they do get propagated to Training jobs .

The other thing we notice is that things like subnets and security group do get propogated from Estimator object to Endpoints and models

• Exact command to reproduce:
  Using sample notebook and modifying :
  notebooks/scikit_learn_inference_pipeline_2019-01-07/Inference Pipeline with Scikit-learn and Linear Learner.ipynb

from sagemaker.sklearn.estimator import SKLearn
script_path = 'sklearn_abalone_featurizer.py'
tagtobeadded = [{'Key':'TagtestKey','Value':'TagtestValue'}]
sklearn_preprocessor = SKLearn(
    entry_point=script_path,
    role=role,
    tags=tagtobeadded,
    train_instance_type="ml.c4.xlarge",
    sagemaker_session=sagemaker_session)

The below training Job will have tags propogated
sklearn_preprocessor.fit({'train': train_input})

The below Endpoint and Model created will not have the tag information

endpoint_name = 'TagsTest'
sklearn_preprocessor.deploy(initial_instance_count=1, instance_type='ml.c4.large', endpoint_name=endpoint_name)

[nadiaya]

Currently tags from training job are not getting propagated to the Endpoint and Model. Adding a feature request label.

As a temporarily work around, while we are getting this change prioritized and implemented, you can try doing the following:

model = sklearn_preprocessor.create_model()
model.deploy(
            instance_type=instance_type,
            initial_instance_count=initial_instance_count,
            endpoint_name=endpoint_name,
            tags=tagtobeadded)

[mvsusp]

Hello @AnshumanRanjan

Thanks for your feature request. We just merged your requested changes to master branch. These changes will be available in the next version released.

Thanks for using SageMaker.

[mvsusp]

Merged additional changes #732

[laurenyu]

#732 and #741 have been released. closing this issue.

[nathansegers]

The propagation of tags is correct now, but the documentation is not clear.
The arguments in the referenced lines read that the tags can be used as a parameter. Yet they cannot be given as argument nor keyword-argument (**kwargs)

It is not clear that the Estimator's tags can be set by using estimator.tags = [{'Key': 'tagname', 'Value': 'tagvalue'}]

sagemaker-python-sdk/src/sagemaker/estimator.py

Lines 351 to 354 in d1e2ab7

	tags(List[dict[str, str]]): Optional. The list of tags to attach to this specific endpoint. Example:
	>>> tags = [{'Key': 'tagname', 'Value': 'tagvalue'}]
	For more information about tags, see https://boto3.amazonaws.com/v1/documentation\
	/api/latest/reference/services/sagemaker.html#SageMaker.Client.add_tags