Repository description

This repo aims to demonstrate the use of AWS lambda function using a Dockerized image. It uses the following tools

• AWS CLI (MakeFile interaction)
• An AWS account with sufficient permissions
• Terraform (IaC)
• Docker (container)

The repo allows you to choose between a python base image provided by aws public.ecr.aws/lambda/python:3.XX or from a custom image of your choice (See Dockerfile and Dockerfile.custom for each respective options)

🔧 Using MakeFile

⚠️ Don't forget to use a .set-env.sh file in the tf/ folder

the whole infrastructure can be deployed with the following make commands

make create-registry

make deploy-lambda

💣 Without the MakeFile

1. Create the ECR and deploy with terraform

in tf/ecr you will find the terraform configuration to deploy the ecr registry. The config will create an output registry_url that can be reused when deploying the lambda function

2. Build and push the docker image

⏩ refer to the official documentation for more information

First login to the created ECR

docker login --username AWS --password-stdin $REPO_URL

Build & Push the image

docker build --no-cache --platform linux/amd64 -t $REPOSITORY_NAME:latest .; \
docker tag $REPOSITORY_NAME:latest $REPO_URL:latest; \
docker push $$REPO_URL:latest

Where:

• $REPOSITORY_NAME is the tag (name) of your docker image
• $REPO_URL is the full url to access your image (<$AWS_ECR_URL/$REPOSITORY_NAME>)

3. Deploy your lambda function

in the tf/lambda folder your will find the information to deploy your lambda function

💁 Terraform will prompt you for registry_url you can fetch the value from 2.

To test

• Use poetry in the Dockerfile, poetry files have been added if you want to use these in the DockerFile

Identified Issues

• Lambda function gets a KMS access denied to read from registry from time to time, deployement is not idempotent w.r.t that error