Serverless Deployment

The demo application can be deployed via AWS Lambda as a serverless web application. The Redis database needs to be accessible from the VPC in which the AWS Lambda function runs. Otherwise, the application remains unchanged.

The Flask application is invoked via AWS Lambda using the flask-serverless library. The application must be packaged into a zip file containing all the dependancies and the invocation function.

You can package the application by:

Configure your environment:
```
export DOCKERID=yourdockerid
export BASE=$HOME/workspace/github
mkdir -p $BASE
cd $BASE
```
where the yourdockerid is your dockerhub identifier and BASE points to the directory where you want your code locally for the build.
If you haven’t already done so, checkout the rediner project:
```
git clone https://github.com/redis-developer/rediner.git
```
Checkout the flask-serverless project as a sibling directory:
```
git clone https://github.com/alexmilowski/flask-serverless.git
```
You should now have the rediner and flask-serverless as sibling directories.
The build script container image guarantees the deployment architecture matches for the packaged python packages. Build this image by:
```
cd rediner/serverless
docker build . -t $DOCKERID/rediner-demo-lambda-build
```

Build the packaging via the docker image:

docker run --rm -v `pwd`:/build -v $BASE/rediner/:/rediner -v $BASE/flask-serverless/:/flask-serverless  $DOCKERID/rediner-demo-lambda-build:latest

After the container exits, you should have a lambda.zip to upload for the AWS Lambda function.

To create the AWS Lambda function:

Choose an Python 3.8 environment
Select the appropriate VPC to enable access to your database
Upload the lambda.zip as the implementation
In “Basic Settings”, change the handler to “production.lambda_handler” and adjust the timeout and memory (e.g., 2 minutes and 256MB)
Set the GRAPH, REDIS_HOST, REDIS_PORT, and REDIS_PASSWORD environment variables to the appropriate values.

To expose the AWS Lambda function as a web application, you will will have to configure an API Gateway and with a Lambda Proxy for the root resource “/” and “/{proxy+}”.