In the many cases, we’re using similar packages to build own Alexa Skill by AWS Lambda.
We have own package.json or npm install commands lists and run them before starting to build the skill.

But for now, we don’t have to do the repeating tasks, because we can easy to import these packages by AWS Lambda Layers.

In this posts, we can learn how to deploy AWS Lambda Layers by Serverless Framework.

Goal

We can launch an AWS Lambda for Alexa without running npm i -S ask-sdk.

Prepare

To begin, we need to install the latest Serverless Framework. Please run npm i -g serverless@latest to install new it.

Make directories

To adding own library into the Layer, we have to place it to the specified place.

Including Library Dependencies in a Layer

Today, We will put a npm library into the Layer. So we’ll create the following directories structures.

$ tree -L 3
.
├── files
│   └── nodejs
│       ├── node_modules
│       ├── package.json
│       └── yarn.lock
└── serverless.yml

And the commands it the following.

$ touch serverless.yml
$ mkdir -p files/nodejs
$ cd files/nodejs
$ npm init -y
$ yarn add ask-sdk ask-utils

Making the serverless.yml

In the Serverless Framework, we can define the Layer like the following example.

service: ask-sdk-layers
frameworkVersion: ">=1.34.0"
provider:
  name: aws
  runtime: nodejs8.10

layers:
  helloLayer:
    description: ASK SDK Packages
    path: files
    compatibleRuntimes:
      - nodejs8.10

The frameworkVersion property is not necessary, but the older serverlesscommand will not work well. So to define it is more kindly for other member.

Deploy

After defining the YAML file, we can deploy by sls deploycommand.

Service Information
service: ask-sdk-layer
stage: dev
region: us-east-1
stack: ask-sdk-layers-dev
api keys:
  None
endpoints:
  None
functions:
  None
layers:
  helloLayer: arn:aws:lambda:us-east-1:99999999:layer:helloLayer:1

You can put the layer into your Lambda Function to put the ARN that shows at layers.

Using layer from the Serverless Framework Project

Finally, we’re trying to use the layer from an AWS Lambda deploying by Serverless Framework. We can put the Layer by following properties.

service: first-alexa
provider:
  name: aws
  runtime: nodejs8.10

functions:
  hello:
    handler: index.handler
    layers:
      - arn:aws:lambda:us-east-1:99999999:layer:helloLayer:4

And if we define the same stack, we can use CloudFormation function like this.

service: practice-of-layer
frameworkVersion: ">=1.34.0"
provider:
  name: aws
  runtime: nodejs8.10

layers:
  hello:
    description: My first layer
    path: files
    compatibleRuntimes:
      - nodejs8.10

functions:
  hello:
    handler: index.handler
    layers:
      - {Ref: HelloLambdaLayer}