Simple and basic example to demonstrate how to run FastAPI on AWS Lambda Serverless environment.
- AWS API Gateway and Lambda for REST API deployment and hosting
- AWS Serverless Application Model for AWS infra deployment
- Fastapi for enabling solutions for all above three challenges
- Mangum for enabling ASGI server on Lambda
- Poetry for Python project and dependency management
- Docker for building AWS Lambda layer
- Python specific standard, asyncronous successor for long lived WSGI
- Intends to provide a standard interface between async-capable Python web servers, frameworks, and applications
- Enables separating protocol specific server and application server concerns in to separate components
NOTE: Precondition is to have Python and Poetry installed.
Install the project dependencies:
poetry install
Run FastAPI application locally:
poetry run uvicorn fastapi_serverless.app:app --reload
Access the interactive docs: http://127.0.0.1:8000/docs
Lambda Layer contains all the python dependencies. Using layer is optional but it's a convenient way to make lambda deployment more efficient.
Run the script to build lambda_layer.zip:
cd scripts
./make_layer.shFollow the instructions on how to create Lambda Layer. Lambda Layer ARN is required in next step.
Project root contains file deploy_TEMPLATE.sh. Update the placeholder attributes (e.g. Lambda Layer ARN) and run the deployment:
cd scripts
./deploy.shIn AWS console navigate to API Gateway and find your newly deployed API (e.g. dev-fastapi-serverless). Under "Stages" find the "Invoke URL".
Access the interactive docs: /docs
Delete SAM deployment:
sam delete --region <AWS_REGION> --stack-name <STACK_NAME>Remember also delete the created AWS Lambda layer.