How to Develop AWS Lambda Functions Locally

LocalStack is the best way to develop AWS Lambda functions.

Most lambda development frameworks, like the Serverless framework or AWS SAM, are all-in-one solutions which tightly couple how you develop and deploy your lambda functions.

All-in-one solutions do not integrate well with other, and arguably better, infrastructure as code tools like Terraform.

But regardless your intentions, LocalStack lets you decouple development from deployment so that you can use the right tool for the right job.

How to Setup LocalStack

LocalStack works by mimicking AWS within localhost managed and ran through Docker Compose.

💻 A fully functional local AWS cloud stack. Develop and test your cloud & Serverless apps offline! (https://github.com/localstack/localstack)

To get started, create a docker-compose.yml file like the one below

version: '3.8'

services:
  localstack:
    image: localstack/localstack:2.0.1
    environment:
      LS_LOG: trace
    ports:
      - 4566:4566
    volumes:
      - /var/run/docker.sock:/var/run/docker.sock

and then run docker-compose up.

Once LocalStack is Ready you can use the AWS CLI to create and invoke your functions locally.

How to Create a Lambda in LocalStack

Just like how you can use AWS CLI to create AWS Lambda functions in production; you can use the same commands to create functions in LocalStack.

To demonstrate how, lets create a “Hello World” lambda.

First, lets create a ./src/hello-world-v0/index.js file which will act as our handler function.

// ./src/hello-world-v0/index.js
exports.handler = async (event = {}, ctx = {}) => {
 console.log({ event })
 return { event }
}

Our “Hello World” lambda will return whatever event we pass to our lambda later.

Now that we have our lambda handler setup, we need to create our function in LocalStack using the aws lambda create-function command.

Creating and running a ./01_create_lambda.bash script like

#!/usr/bin/env bash
# ./01_create_lambda.bash
set -e

function create_lambda() {
 # ! Absolute path to code
 local dir=$PWD/src/hello-world-v0/

 # * LocalStack localhost endpoint
 local endpoint=http://localhost:4566

 # * Lambda configuration
 local function_handler=index.handler
 local function_name=hello-world-v0
 local function_role=arn:aws:iam::000000000000:role/localstack-does-not-care
 local function_runtime=nodejs18.x

 aws --endpoint-url $endpoint lambda delete-function --function-name $function_name || true

 # * Create lambda in LocalStack
 aws --endpoint-url $endpoint lambda create-function \
  --code S3Bucket="hot-reload",S3Key="$dir" \
  --function-name $function_name \
  --handler $function_handler \
  --role $function_role \
  --runtime $function_runtime
}

create_lambda

we will create our “Hello World” function in LocalStack.

How to Invoke a Lambda in LocalStack

Just like creating a lambda function with the AWS CLI, you can also invoke a lambda using the AWS CLI.

Creating and running a ./02_invoke_lambda.bash script to invoke our LocalStack lambda with a payload of { "hello": "world" } like

#!/usr/bin/env bash
# ./02_invoke_lambda.bash
set -e

function invoke_lambda() {
 # * LocalStack localhost endpoint
 local endpoint=http://localhost:4566

 # * Lambda configuration
 local function_name=hello-world-v0

 # * Invoke configuration
 local invoke_payload='{ "hello": "world" }'
 local invoke_to_file=response.json
 local invoke_type=RequestResponse

 # * Invoke lambda from LocalStack
 aws --endpoint $endpoint lambda invoke \
  --cli-binary-format raw-in-base64-out \
  --function-name $function_name \
  --invocation-type $invoke_type \
  --payload "$invoke_payload" \
  $invoke_to_file
}

invoke_lambda

we should get a response of {"event": { "hello": "world"} } in our ./response.json file.

Conclusion

LocalStack works by mimicking AWS on y our local computer.

This lets you create, invoke, and delete lambda functions as if they were in a production environment.

This also decouples your development process from your deployment process so that you can overall managed your lambda functions however you see fit.

References

• Complete GitHub example