Create, Store, and Share Features
Last updated
Was this helpful?
Last updated
Was this helpful?
Important Use the latest version of Amazon SageMaker Studio to make sure that you're using the most recent version of the search functionality. For information on updating Studio, see .
Note The feature groups that you're searching for must be within the same AWS account and AWS Region.
Important Use the latest version of Amazon SageMaker Studio to make sure that you're using the most recent version of the search functionality. For information on updating Studio, see .
Note
The role that you use must have the following managed policies attached to it: AmazonS3FullAccess
and AmazonSageMakerFeatureStoreAccess
.
Important Feature Store requires data to be registered in a AWS Glue data catalog. By default, Feature Store automatically builds an AWS Glue data catalog when you create a feature group.
Note To make sure that your data is up to date, you can set up a AWS Glue crawler to run on a schedule. To set up a AWS Glue crawler, specify an IAM role that the crawler is using to access the offline store’s S3 buckets. For more information, see . For more information on how to use AWS Glue and Athena to build a training dataset for model training and inference, see .
Note
The key policy for the online store also works for the offline store, only when the kms:ViaService
condition is not specified.
Note Soft limits can be increased based on your needs.
Important You can specify a AWS KMS encryption key to encrypt the Amazon S3 location used for your offline feature store when you create a feature group. If AWS KMS encryption key is not specified, by default we encrypt all data at rest using AWS KMS key. By defining your for SSE, you can reduce AWS KMS requests costs by up to 99 percent.