Creating Aurora zero-ETL integrations with an Amazon SageMaker lakehouse - Amazon Aurora
PrerequisitesRequired permissionsCreating zero-ETL integrations with an Amazon SageMaker lakehouseEncrypting integrationsNext steps

Creating Aurora zero-ETL integrations with an Amazon SageMaker lakehouse

When you create an Aurora zero-ETL integration with an Amazon SageMaker lakehouse, you specify the source Aurora DB cluster and the target AWS Glue managed catalog. You can also customize encryption settings and add tags. Aurora creates an integration between the source DB cluster and its target. Once the integration is active, any data that you insert into the source DB cluster will be replicated into the configured target.

Prerequisites

Before you create a zero-ETL integration with an Amazon SageMaker lakehouse, you must create a source DB cluster and a target AWS Glue managed catalog. You also must allow replication into the catalog by adding the DB cluster as an authorized integration source.

For instructions to complete each of these steps, see Getting started with Aurora zero-ETL integrations.

Required permissions

Certain IAM permissions are required to create a zero-ETL integration with an Amazon SageMaker lakehouse. At minimum, you need permissions to perform the following actions:

  • Create zero-ETL integrations for the source Aurora DB cluster.

  • View and delete all zero-ETL integrations.

  • Create inbound integrations into the target AWS Glue managed catalog.

  • Access Amazon S3 buckets used by the AWS Glue managed catalog.

  • Use AWS KMS keys for encryption if custom encryption is configured.

  • Register resources with Lake Formation.

  • Put resource policy on the AWS Glue managed catalog to authorize inbound integrations.

The following sample policy demonstrates the least privilege permissions required to create and manage integrations with an Amazon SageMaker lakehouse. You might not need these exact permissions if your user or role has broader permissions, such as an AdministratorAccess managed policy.

Additionally, you must configure a resource policy on the target AWS Glue managed catalog to authorize inbound integrations. Use the following AWS CLI command to apply the resource policy.

aws glue put-resource-policy \
      --policy-in-json  '{
    "Version": "2012-10-17",
    "Statement": [{
        "Effect": "Allow",
        "Principal": {
            "Service": "glue.amazonaws.com"
        },
        "Action": [
            "glue:AuthorizeInboundIntegration"
        ],
        "Resource": ["arn:aws:glue:region:account_id:catalog/catalog_name"],
        "Condition": {
            "StringEquals": {
                "aws:SourceArn": "arn:aws:rds:region:account_id:db:source_name"
            }
        }
    },
    {
        "Effect": "Allow",
        "Principal": {
            "AWS": "account_id"
        },
        "Action": ["glue:CreateInboundIntegration"],
        "Resource": ["arn:aws:glue:region:account_id:catalog/catalog_name"]
    }
    ]
}' \
      --region region
Note

Glue catalog Amazon Resource Names (ARNs) have the following format:

  • Glue catalog – arn:aws:glue:{region}:{account-id}:catalog/catalog-name

Choosing a target AWS Glue managed catalog in a different account

If you plan to specify a target AWS Glue managed catalog that's in another AWS account, you must create a role that allows users in the current account to access resources in the target account. For more information, see Providing access to an IAM user in another AWS account that you own.

The role must have the following permissions, which allow the user to view available AWS Glue catalogs in the target account.

{
   "Version":"2012-10-17",
   "Statement":[
      {
         "Effect":"Allow",
         "Action":[
            "glue:GetCatalog"
         ],
         "Resource":[
            "*"
         ]
      }
   ]
}

The role must have the following trust policy, which specifies the target account ID.

{
   "Version":"2012-10-17",
   "Statement":[
      {
         "Effect":"Allow",
         "Principal":{
            "AWS": "arn:aws:iam::{external-account-id}:root"
         },
         "Action":"sts:AssumeRole"
      }
   ]
}

For instructions to create the role, see Creating a role using custom trust policies.

Creating zero-ETL integrations with an Amazon SageMaker lakehouse

You can create a zero-ETL integration with an Amazon SageMaker lakehouse using the AWS Management Console, the AWS CLI, or the RDS API.

Important

Zero-ETL integrations with an Amazon SageMaker lakehouse do not support refresh or resync operations. If you encounter issues with an integration after creation, you must delete the integration and create a new one.

To create a zero-ETL integration with an Amazon SageMaker lakehouse

  1. Sign in to the AWS Management Console and open the Amazon RDS console at https://console.aws.amazon.com/rds/.

  2. In the left navigation pane, choose Zero-ETL integrations.

  3. Choose Create zero-ETL integration.

  4. For Integration identifier, enter a name for the integration. The name can have up to 63 alphanumeric characters and can include hyphens.

  5. Choose Next.

  6. For Source, select the Aurora DB cluster where the data will originate from.

    Note

    RDS notifies you if the DB cluster parameters aren't configured correctly. If you receive this message, you can either choose Fix it for me, or configure them manually. For instructions to fix them manually, see Step 1: Create a custom DB cluster parameter group.

    Modifying DB cluster parameters requires a reboot. Before you can create the integration, the reboot must be complete and the new parameter values must be successfully applied to the cluster.

  7. (Optional) Select Customize data filtering options and add data filters to your integration. You can use data filters to define the scope of replication to the target Amazon SageMaker lakehouse. For more information, see Data filtering for Aurora zero-ETL integrations.

  8. Once your source DB cluster is successfully configured, choose Next.

  9. For Target, do the following:

    1. (Optional) To use a different AWS account for the Amazon SageMaker lakehouse target, choose Specify a different account. Then, enter the ARN of an IAM role with permissions to display your AWS Glue catalogs. For instructions to create the IAM role, see Choosing a target AWS Glue managed catalog in a different account.

    2. For AWS Glue catalog, select the target for replicated data from the source DB cluster. You can choose an existing AWS Glue managed catalog as the target.

    3. The target IAM role needs describe permissions on the target catalog and must have the following permissions:

      {
    "Version": "2012-10-17",
    "Statement": [
        {
            "Sid": "VisualEditor0",
            "Effect": "Allow",
            "Action": "glue:GetCatalog",
            "Resource": [
                "arn:aws:glue:region:account-id:catalog/*",
                "arn:aws:glue:region:account-id:catalog"
            ]
        }
    ]
}

      The target IAM role must have the following trust relationship:

      {
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Principal": {
                "Service": "glue.amazonaws.com"
            },
            "Action": "sts:AssumeRole"
        }
    ]
}

    4. You must grant the target IAM role describe permissions for the target AWS Glue managed catalog with the Lake Formation administrator role created in Step 3b: Create an AWS Glue catalog for Amazon SageMaker AI zero-ETL integration.

    Note

    RDS notifies you if the resource policy or configuration settings for the specified AWS Glue managed catalog aren't configured correctly. If you receive this message, you can either choose Fix it for me, or configure them manually.

    If your selected source and target are in different AWS accounts, then Amazon RDS cannot fix these settings for you. You must navigate to the other account and fix them manually in SageMaker Unified Studio.

  10. Once your target AWS Glue managed catalog is configured correctly, choose Next.

  11. (Optional) For Tags, add one or more tags to the integration. For more information, see Tagging Amazon Aurora and Amazon RDS resources.

  12. For Encryption, specify how you want your integration to be encrypted. By default, RDS encrypts all integrations with an AWS owned key. To choose a customer managed key instead, enable Customize encryption settings and choose a KMS key to use for encryption. For more information, see Encrypting Amazon Aurora resources.

    Optionally, add an encryption context. For more information, see Encryption context in the AWS Key Management Service Developer Guide.

    Note

    Amazon RDS adds the following encryption context pairs in addition to any that you add:

    • aws:glue:integration:arn - IntegrationArn

    • aws:servicename:id - glue

    This reduces the overall number of pairs that you can add from 8 to 6, and contributes to the overall character limit of the grant constraint. For more information, see Using grant constraints in the AWS Key Management Service Developer Guide.

  13. Choose Next.

  14. Review your integration settings and choose Create zero-ETL integration.

    If creation fails, see Troubleshooting Aurora zero-ETL integrations for troubleshooting steps.

The integration has a status of Creating while it's being created, and the target Amazon SageMaker lakehouse has a status of Modifying. During this time, you can't query the catalog or make any configuration changes on it.

When the integration is successfully created, the status of the integration and the target Amazon SageMaker lakehouse both change to Active.

To prepare a target AWS Glue managed catalog for zero-ETL integration using the AWS CLI, you must first use the create-integration-resource-property command with the following options:

  • --resource-arn – Specify the ARN of the AWS Glue managed catalog that will be the target for the integration.

  • --target-processing-properties – Specify the ARN of the IAM role to access the target AWS Glue managed catalog 

aws glue create-integration-resource-property --region us-east-1
 --resource-arn arn:aws:glue:region:account_id:catalog/catalog_name \
 --target-processing-properties '{"RoleArn" : "arn:aws:iam::account_id:role/TargetIamRole"}'

To create a zero-ETL integration with an Amazon SageMaker lakehouse using the AWS CLI, use the create-integration command with the following options:

  • --integration-name – Specify a name for the integration.

  • --source-arn – Specify the ARN of the Aurora DB cluster that will be the source for the integration.

  • --target-arn – Specify the ARN of the AWS Glue managed catalog that will be the target for the integration.

For Linux, macOS, or Unix:

aws rds create-integration \
    --integration-name my-sagemaker-integration \
    --source-arn arn:aws:rds:{region}:{account-id}:my-db \
    --target-arn arn:aws:glue:{region}:{account-id}:catalog/catalog-name

For Windows:

aws rds create-integration ^
    --integration-name my-sagemaker-integration ^
    --source-arn arn:aws:rds:{region}:{account-id}:my-db ^
    --target-arn arn:aws:glue:{region}:{account-id}:catalog/catalog-name

To create a zero-ETL integration with Amazon SageMaker by using the Amazon RDS API, use the CreateIntegration operation with the following parameters:

Note

Catalog names are limited to 19 characters. Ensure your IntegrationName parameter meets this requirement if it will be used as a catalog name.

  • IntegrationName – Specify a name for the integration.

  • SourceArn – Specify the ARN of the Aurora DB cluster that will be the source for the integration.

  • TargetArn – Specify the ARN of the AWS Glue managed catalog that will be the target for the integration.

Encrypting integrations with a customer managed key

If you specify a custom KMS key rather than an AWS owned key when you create an integration with Amazon SageMaker, the key policy must provide the SageMaker Unified Studio service principal access to the CreateGrant action. In addition, it must allow the current user to perform to the DescribeKey and CreateGrant actions.

The following sample policy demonstrates how to provide the required permissions in the key policy. It includes context keys to further reduce the scope of permissions.

{
    "Version": "2012-10-17",
    "Id": "Key policy",
    "Statement": [
        {
            "Sid": "Enables IAM user permissions",
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::{account-ID}:root"
            },
            "Action": "kms:*",
            "Resource": "*"
        },
        {
            "Sid": "Allows the Glue service principal to add a grant to an AWS KMS key",
            "Effect": "Allow",
            "Principal": {
                "Service": "glue.amazonaws.com"
            },
            "Action": "kms:CreateGrant",
            "Resource": "*",
            "Condition": {
                "StringEquals": {
                    "kms:EncryptionContext:{context-key}":"{context-value}"
                },
                "ForAllValues:StringEquals": {
                    "kms:GrantOperations": [
                        "Decrypt",
                        "GenerateDataKey",
                        "CreateGrant"
                    ]
                }
            }
        },
        {
            "Sid": "Allows the current user or role to add a grant to a KMS key",
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::{account-ID}:role/{role-name}"
            },
            "Action": "kms:CreateGrant",
            "Resource": "*",
            "Condition": {
                "StringEquals": {
                    "kms:EncryptionContext:{context-key}":"{context-value}",
                    "kms:ViaService": "rds.us-east-1.amazonaws.com"
                },
                "ForAllValues:StringEquals": {
                    "kms:GrantOperations": [
                        "Decrypt",
                        "GenerateDataKey",
                        "CreateGrant"
                    ]
                }
            }
        },
        {
            "Sid": "Allows the current uer or role to retrieve information about a KMS key",
            "Effect": "Allow",
            "Principal": {
                "AWS": "arn:aws:iam::{account-ID}:role/{role-name}"
            },
            "Action": "kms:DescribeKey",
            "Resource": "*"
        }
    ]
}

For more information, see Creating a key policy in the AWS Key Management Service Developer Guide.

Next steps

After you successfully create a zero-ETL integration with Amazon SageMaker, you can start adding data to the source Aurora DB cluster and querying it in your Amazon SageMaker lakehouse. The data will be automatically replicated and made available for analytics and machine learning workloads.