Creating a common library for all functions makes code:

• more efficient
• readable
• predictable
• testable

Particularly when validating input and creating common response formats for all functions. I have a bad habit of inconsistent response formats that break other parsers 😛 Creating a common format and passing arguments from each function is much more maintainable

I ran into some headaches when my libraries required the installation of specific packages. Just sticking requirements.txt in the folder was not sufficient for SAM’s processes to automatically include them. Some extra elements are required in the tempalte.yaml file

To create a library of functions available across all lambda functions requires:

1. A separate folder with the functions
2. its own requirements file
3. definitions in SAM

Folder Structure

Project
|-.aws-sam
|-layers
| |-common
| | |-helpers.py
| | |-validator.py
| | |-requirements.txt
| |-special
| | |-quantum_entangler.py
| | |-requirements.txt
|-foo_function
| |-app.py
| |-requirements.txt
|-bar_function
| |-app.py
| |-requirements.txt
|-.gitignore
|-template.yaml

Layer code does not need to be stored inside a folder named dependencies\\python . This is only required if you are building your own layer and providing a .zip file in the template.

template.yaml

Globals:            # Common to all lambda resources in the stack
  Function:
    Layers:
    - !Ref CommonFunctionsLayer

Resources:
	CommonFunctionsLayer:
    Type: AWS::Serverless::LayerVersion
    Properties:
      LayerName: sam-app-dependencies
      Description: Common helper functions shared across Lambda functions
      ContentUri: layers/common/
      CompatibleRuntimes:
        - python3.8
      RetentionPolicy: Delete
    Metadata:                       # THIS IS THE IMPORTANT BIT!
      BuildMethod: python3.8
      BuildArchitecture: x86_64 # arm64 is an alternative and recommended for performance (I suspect there may be compatability problems)
	
	SpecialFunctionLayer:
    Type: AWS::Serverless::LayerVersion
    Properties:
      LayerName: special_layer
      Description: Special helper function used only in one function
      CompatibleRuntimes:
        - python3.8
      RetentionPolicy: Delete
    Metadata:
      BuildMethod: python3.8
      BuildArchitecture: x86_64 

	FooFunction:
    Type: AWS::Serverless::Function # More info about Function Resource: <https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#awsserverlessfunction>
    Description: this is some generic function
    Properties:
      CodeUri: foo_functioon/
      Handler: app.lambda_handler
      Runtime: python3.8
      Events:
        ApiEvent:
          Type: Api # More info about API Event Source: <https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api>
          Properties:
            Path: /foo
            Method: put
	
BarFunction:
    Type: AWS::Serverless::Function # More info about Function Resource: <https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#awsserverlessfunction>
    Description: this is some generic function
    Properties:
      CodeUri: bar_functioon/
      Handler: app.lambda_handler
      Runtime: python3.8
      Events:
        ApiEvent:
          Type: Api # More info about API Event Source: <https://github.com/awslabs/serverless-application-model/blob/master/versions/2016-10-31.md#api>
          Properties:
            Path: /foo
            Method: put
		Layers:
    - !Ref SpecialFunctionLayer

Both metadata parameters must be defined for SAMs processes to build the layers from scratch