Easily deploy any Node.js web server to AWS Lambda.
Scandium can deploy your Express, Koa or similar app to lambda, without any modifications to your code. Combined with AWS API Gateway you can have your current API running serverless in no time 😎
Scandium requires Node.js, Docker, and a working AWS config w/ credentials. If you are unfamiliar with the AWS config an easy way to get started is to install the AWS CLI and run aws configure
.
macOS w/ Brew: brew install node
brew cask install docker
brew install awscli
aws configure |
Linux w/ Snap: sudo snap install node --classic
sudo snap install docker
sudo snap install aws-cli --classic
aws configure |
Windows w/ winget: winget install OpenJS.NodeJS
winget install Docker.DockerDesktop
winget install Amazon.AWSCLI
aws configure |
npm install --global scandium
To create a new Lambda function, use the scandium create
command. This will package your application and upload it to AWS Lambda, as well as configure an API Gateway in front of your function.
scandium create --name=my-awesome-api
You should now be presented with a url where you can access your api.
Now serving live requests at: https://xxxxxxxxxx.execute-api.us-west-2.amazonaws.com/default
Whenever you make changes to your app, you can upload a new version of it by using the scandium update
command. This will package your application again, and upload it as a new version to the specified Lambda function. It will also update the configuration of the API Gateway to point at the new function.
scandium update --name=my-awesome-api
By default, Scandium will set up an API Gateway that simply forwards all requests to the Lambda function. If you want to utilise the benefits of API Gateway fully, you can provide a Swagger file describing your API endpoints. Pass the --swagger=my-api-definition.yml
to either the create
or update
command and Scandium will configure the API Gateway for you.
Scandium has support for prepare
/build
scripts, if the script is present in the package.json
it will make sure that the script is being run with full devDependencies
installed. The final package being uploaded to Lambda will still only contain the production dependencies
.
If both a prepare
and a build
script is present, Scandium will pick the prepare
script.
If there is a .dockerignore
file present, that one will be used when building the app. Otherwise, if a .gitignore
file is present, that one will be used. If none of these files exists, a built in list that just contains .git
and node_modules
will be used.
If you want to run any specific script inside the Lambda once during each deploy (e.g. for running database migrations), you can use the --hooks
flag. Pass it the name of a file exporting functions for each hook you want to run.
Currently there is only one hook, named deploy
, that will run after your Lambda is created, and before the API Gateway is updated to the new Lambda.
Example file:
const runDatabaseMigrations = require('...')
exports.deploy = async function () {
await runDatabaseMigrations()
}