@aerisweather/deploy-lambda-function

Helper for deploying code to AWS Lambda.

What does it do

• Packages your code into a zip file (excludes devDependencies, to keep it small)
• Uploads your code to a lambda function
• Loads env vars from S3, and applies them to your new lambda function version
• Sets the IAM role of your new lambda function version
• Points an alias at your new lambda function version

Install

Install via npm:

npm install --save-dev @aerisweather/deploy-lambda-function

Usage

First, you'll need to setup a param.json config files for each deployment environment. For example:

// params.staging.json
{
  // List of directories to include in `zip` file
  "srcDirs": ["dist/lib", "node_modules"],
  // Name of your lambda function on AWS
  "lambdaFunction": "my-lambda-function",
  // Location of environment variables on S3
  "envFile": "s3://my-bucket/env/staging.env",
  // The alias you want to point at your lambda function
  "lambdaAlias": "staging",
  // The IAM role you want to give your lambda function
  "lambdaRole": "arn:aws:iam::ACCOUNT_ID:role/my-lambda-function-staging",
  // The region associated with your lambda function (optional, defaults to us-east-1)
  "lambdaRegion": "us-east-1"
}

To deploy to that environment, run:

node_modules/.bin/deploy-lambda-function -c ./params.staging.json

Programmatic Usage

You can also use this library programmatically, via a node script:

const deployLambdaFunction = require('@aerisweather/deploy-lambda-function').default;

deployLambdaFunction({
  srcDirs: ['dist/lib', 'node_modules'],
  lambdaFunction: 'my-lambda-function',
  envFile: 's3://my-bucket/env/staging.env',
  lambdaAlias: 'staging',
  lambdaRole: 'arn:aws:iam::ACCOUNT_ID:role/my-lambda-function-staging',
  lambdaRegion: 'us-east-1'
})

Example: BitBucket Pipelines

You can setup BitBucket pipelines to deploy to lambda, using deploy-lambda-function.

For this to work, you'll need to:

• Create params.json files for prod, staging, and dev
• Add the bitbucket-pipelines.yml file below to your repo

# bitbucket-pipelines.yml

# Run on LambCI's lambda-simulation container,
# to give an environment that's close to the real deal
image: lambci/build-nodejs4.3

clone:
  depth: 1

pipelines:
  # Default build scripts...
  default:
    - step:
        script:
          # Configure npm with your NPM_TOKEN
          # (only required if installing from a private npm registry)
          - touch .npmrc && echo "//registry.npmjs.org/:_authToken=${NPM_TOKEN}" >> .npmrc
          # Build and test your code
          - npm install -g npm
          - npm install
          - npm run build
          - npm run test

  # master/staging/development get deployed to lambda
  branches:
    master:
      - step:
          script:
            # Configure npm with your NPM_TOKEN
            # (only required if installing from a private npm registry)
            - touch .npmrc && echo "//registry.npmjs.org/:_authToken=${NPM_TOKEN}" >> .npmrc
            # Build your code
            - npm install && npm build && npm test
            # Deploy to lambda, using prod config
            - ./node_modules/.bin/deploy-lambda-function -c ./deploy/params.prod.json
    staging:
      - step:
          script:
            # Configure npm with your NPM_TOKEN
            # (only required if installing from a private npm registry)
            - touch .npmrc && echo "//registry.npmjs.org/:_authToken=${NPM_TOKEN}" >> .npmrc
            # Build your code
            - npm install && npm build && npm test
            # Deploy to lambda, using staging config
            - ./node_modules/.bin/deploy-lambda-function -c ./deploy/params.staging.json
    development:
      - step:
          script:
            # Configure npm with your NPM_TOKEN
            # (only required if installing from a private npm registry)
            - touch .npmrc && echo "//registry.npmjs.org/:_authToken=${NPM_TOKEN}" >> .npmrc
            # Build your code
            - npm install && npm build && npm test
            # Deploy to lambda, using dev config
            - ./node_modules/.bin/deploy-lambda-function -c ./deploy/params.dev.json

• Create an IAM user for Pipelines, with an IAM policy like so

{
    "Version": "2012-10-17",
    "Statement": [
        {
            "Effect": "Allow",
            "Action": [
                "lambda:PublishVersion",
                "lambda:UpdateAlias",
                "lambda:UpdateFunctionCode",
                "lambda:UpdateFunctionConfiguration"
            ],
            "Resource": [
                "arn:aws:lambda:ACCOUNT_REGION:ACCOUNT_ID:function:NAME_OF_YOUR_LAMBDA_FUNCTION"
            ]
        }
    ]
}

• Assign the IAM role to an IAM user
• Provide IAM user credentials to Pipelines, via env vars
• Provide npm credentials to Pipelines, by copying the token from your local ~/.npmrc to the NPM_TOKEN env var in Pipelines
  • (only required if installing from a private npm registry)