AWS Secrets Manager Construct Library

---

AWS CDK v1 has reached End-of-Support on 2023-06-01. This package is no longer being updated, and users should migrate to AWS CDK v2.

For more information on how to migrate, see the Migrating to AWS CDK v2 guide.

---

import aws_cdk.aws_secretsmanager as secretsmanager

Create a new Secret in a Stack

In order to have SecretsManager generate a new secret value automatically, you can get started with the following:

# Default secret
secret = secretsmanager.Secret(self, "Secret")
# Using the default secret
iam.User(self, "User",
    password=secret.secret_value
)
# Templated secret
templated_secret = secretsmanager.Secret(self, "TemplatedSecret",
    generate_secret_string=secretsmanager.SecretStringGenerator(
        secret_string_template=JSON.stringify({"username": "user"}),
        generate_string_key="password"
    )
)
# Using the templated secret
iam.User(self, "OtherUser",
    user_name=templated_secret.secret_value_from_json("username").to_string(),
    password=templated_secret.secret_value_from_json("password")
)

If you need to use a pre-existing secret, the recommended way is to manually provision the secret in AWS SecretsManager and use the Secret.fromSecretArn or Secret.fromSecretAttributes method to make it available in your CDK Application:

# encryption_key: kms.Key

secret = secretsmanager.Secret.from_secret_attributes(self, "ImportedSecret",
    secret_arn="arn:aws:secretsmanager:<region>:<account-id-number>:secret:<secret-name>-<random-6-characters>",
    # If the secret is encrypted using a KMS-hosted CMK, either import or reference that key:
    encryption_key=encryption_key
)

SecretsManager secret values can only be used in select set of properties. For the list of properties, see the CloudFormation Dynamic References documentation.

A secret can set RemovalPolicy. If it set to RETAIN, that removing a secret will fail.

Grant permission to use the secret to a role

You must grant permission to a resource for that resource to be allowed to use a secret. This can be achieved with the Secret.grantRead and/or Secret.grantWrite method, depending on your need:

role = iam.Role(self, "SomeRole", assumed_by=iam.AccountRootPrincipal())
secret = secretsmanager.Secret(self, "Secret")
secret.grant_read(role)
secret.grant_write(role)

If, as in the following example, your secret was created with a KMS key:

# role: iam.Role

key = kms.Key(self, "KMS")
secret = secretsmanager.Secret(self, "Secret", encryption_key=key)
secret.grant_read(role)
secret.grant_write(role)

then Secret.grantRead and Secret.grantWrite will also grant the role the relevant encrypt and decrypt permissions to the KMS key through the SecretsManager service principal.

The principal is automatically added to Secret resource policy and KMS Key policy for cross account access:

other_account = iam.AccountPrincipal("1234")
key = kms.Key(self, "KMS")
secret = secretsmanager.Secret(self, "Secret", encryption_key=key)
secret.grant_read(other_account)

Rotating a Secret

Using a Custom Lambda Function

A rotation schedule can be added to a Secret using a custom Lambda function:

import aws_cdk.aws_lambda as lambda_

# fn: lambda.Function

secret = secretsmanager.Secret(self, "Secret")

secret.add_rotation_schedule("RotationSchedule",
    rotation_lambda=fn,
    automatically_after=Duration.days(15)
)

Note: The required permissions for Lambda to call SecretsManager and the other way round are automatically granted based on AWS Documentation as long as the Lambda is not imported.

See Overview of the Lambda Rotation Function on how to implement a Lambda Rotation Function.

Using a Hosted Lambda Function

Use the hostedRotation prop to rotate a secret with a hosted Lambda function:

secret = secretsmanager.Secret(self, "Secret")

secret.add_rotation_schedule("RotationSchedule",
    hosted_rotation=secretsmanager.HostedRotation.mysql_single_user()
)

Hosted rotation is available for secrets representing credentials for MySQL, PostgreSQL, Oracle, MariaDB, SQLServer, Redshift and MongoDB (both for the single and multi user schemes).

When deployed in a VPC, the hosted rotation implements ec2.IConnectable:

# my_vpc: ec2.Vpc
# db_connections: ec2.Connections
# secret: secretsmanager.Secret


my_hosted_rotation = secretsmanager.HostedRotation.mysql_single_user(vpc=my_vpc)
secret.add_rotation_schedule("RotationSchedule", hosted_rotation=my_hosted_rotation)
db_connections.allow_default_port_from(my_hosted_rotation)

See also Automating secret creation in AWS CloudFormation.

Rotating database credentials

Define a SecretRotation to rotate database credentials:

# my_secret: secretsmanager.Secret
# my_database: ec2.IConnectable
# my_vpc: ec2.Vpc


secretsmanager.SecretRotation(self, "SecretRotation",
    application=secretsmanager.SecretRotationApplication.MYSQL_ROTATION_SINGLE_USER,  # MySQL single user scheme
    secret=my_secret,
    target=my_database,  # a Connectable
    vpc=my_vpc,  # The VPC where the secret rotation application will be deployed
    exclude_characters=" %+:;{}"
)

The secret must be a JSON string with the following format:

{
  "engine": "<required: database engine>",
  "host": "<required: instance host name>",
  "username": "<required: username>",
  "password": "<required: password>",
  "dbname": "<optional: database name>",
  "port": "<optional: if not specified, default port will be used>",
  "masterarn": "<required for multi user rotation: the arn of the master secret which will be used to create users/change passwords>"
}

For the multi user scheme, a masterSecret must be specified:

# my_user_secret: secretsmanager.Secret
# my_master_secret: secretsmanager.Secret
# my_database: ec2.IConnectable
# my_vpc: ec2.Vpc


secretsmanager.SecretRotation(self, "SecretRotation",
    application=secretsmanager.SecretRotationApplication.MYSQL_ROTATION_MULTI_USER,
    secret=my_user_secret,  # The secret that will be rotated
    master_secret=my_master_secret,  # The secret used for the rotation
    target=my_database,
    vpc=my_vpc
)

See also aws-rds where credentials generation and rotation is integrated.

Importing Secrets

Existing secrets can be imported by ARN, name, and other attributes (including the KMS key used to encrypt the secret). Secrets imported by name should use the short-form of the name (without the SecretsManager-provided suffx); the secret name must exist in the same account and region as the stack. Importing by name makes it easier to reference secrets created in different regions, each with their own suffix and ARN.

secret_complete_arn = "arn:aws:secretsmanager:eu-west-1:111111111111:secret:MySecret-f3gDy9"
secret_partial_arn = "arn:aws:secretsmanager:eu-west-1:111111111111:secret:MySecret" # No Secrets Manager suffix
encryption_key = kms.Key.from_key_arn(self, "MyEncKey", "arn:aws:kms:eu-west-1:111111111111:key/21c4b39b-fde2-4273-9ac0-d9bb5c0d0030")
my_secret_from_complete_arn = secretsmanager.Secret.from_secret_complete_arn(self, "SecretFromCompleteArn", secret_complete_arn)
my_secret_from_partial_arn = secretsmanager.Secret.from_secret_partial_arn(self, "SecretFromPartialArn", secret_partial_arn)
my_secret_from_name = secretsmanager.Secret.from_secret_name_v2(self, "SecretFromName", "MySecret")
my_secret_from_attrs = secretsmanager.Secret.from_secret_attributes(self, "SecretFromAttributes",
    secret_complete_arn=secret_complete_arn,
    encryption_key=encryption_key
)

Replicating secrets

Secrets can be replicated to multiple regions by specifying replicaRegions:

# my_key: kms.Key

secretsmanager.Secret(self, "Secret",
    replica_regions=[secretsmanager.ReplicaRegion(
        region="eu-west-1"
    ), secretsmanager.ReplicaRegion(
        region="eu-central-1",
        encryption_key=my_key
    )
    ]
)

Alternatively, use addReplicaRegion():

secret = secretsmanager.Secret(self, "Secret")
secret.add_replica_region("eu-west-1")