Sharing Lambda Layers - AWS Serverless Application Repository
How It WorksExample

Sharing Lambda Layers

If you've implemented functionality in a Lambda layer, you might want to share your layer without hosting a global instance of it. Sharing layers in this manner enables others to deploy an instance of your layer to their own account. This prevents client applications from depending on a global instance of your layer. The AWS Serverless Application Repository enables you to share Lambda layers in this manner easily.

For more information about Lambda layers, see AWS Lambda Layers in the AWS Lambda Developer Guide.

How It Works

The following are the steps for sharing your layer using the AWS Serverless Application Repository. This allows a copy of your layer to be created in the user's AWS account.

  1. Define a serverless application with an AWS SAM template that includes your layer as a resource— that is, either an AWS::Serverless::LayerVersion or an AWS::Lambda::LayerVersion resource.

  2. Publish your application to the AWS Serverless Application Repository, and share it (either publicly or privately).

  3. A customer deploys your application, which creates a copy of your layer in their own AWS account. The customer can now reference the Amazon Resource Name (ARN) of the layer in their AWS account in their client application.

Example

The following is an example AWS SAM template for an application that contains the Lambda layer that you want to share:

Resources:
  SharedLayer:
    Type: AWS::Serverless::LayerVersion
    Properties:
      LayerName: shared-layer
      ContentUri: source/layer-code/
      CompatibleRuntimes:
        - python3.7
Outputs:
  LayerArn:
    Value: !Ref SharedLayer

When a customer deploys your application from the AWS Serverless Application Repository, a layer is created in their AWS account. The ARN of the layer looks something like the following:

arn:aws:lambda:us-east-1:012345678901:layer:shared-layer:1

The customer can now reference this ARN in their own client application, like in this example:

Resources:
  MyFunction:
    Type: AWS::Serverless::Function
    Properties:
      Handler: index.handler
      Runtime: python3.7
      CodeUrl: source/app-code/
      Layers:
        - arn:aws:lambda:us-east-1:012345678901:layer:shared-layer:1