Deploy a Serverless API in 30 mins

Presented in Ep05 of the Coffee & Cloud Series

Contact: https://linktr.ee/awsukstartups

Objective

In this example, we will deploy a serverless Python function to API Gateway. The function will interact with a DynamoDB table for persistence.

The scenario is that we are building an dashboard app to show times when a train leaves a station heading to a destination. Each time a train departs we want to update our app to show the destination and the time that it departed.

Requirements

AWS Account
AWS CLI installed & configured
AWS CDK installed & configured
Git CLI installed
Docker installed
Curl or Httpie installed (for testing)

Expected Time

You should be able to deploy this in less than 30 mins.

Getting Started

1. Initiate a new CDK application:

$ cdk init app --language typescript sample-api
$ cd sample-api

2. Add the CDK resources for the application

In the generated lib/sample-api-stack.ts add the DynamoDB Table, Lambda function & the API Gateway definition and allow the function to do read writes on the DynamoDB table.

    const table = new Table(this, 'Table', {
      partitionKey: { type: AttributeType.STRING, name: 'id' },
    })

    const func = new PythonFunction(this, 'MyFunction', {
      entry: './lambda-api',
      runtime: Runtime.PYTHON_3_9,
      index: 'server.py',
      handler: 'handler',
      environment: {
        'TABLE_NAME': table.tableName
      }
    })

    new LambdaRestApi(this, 'RestAPI', {
      handler: func
    })

    table.grantReadWriteData(func)

Note: For the full code see lib/sample-api-stack.ts

Note here we are using the PythonFunction construct here. This is a special construct that can perform a number of Python-based operations on your behalf. In this case, we want to include some Python libraries into our application.

The PythonFunction construct requires that you have docker installed and accessible before you can build and deploy your project.

3. Create your lambda function

Before we can deploy the CDK stack, we need to have the code ready for the function. A sample Python function has been provided in lambda-api/.

Copy the folder lambda-api/ to your application. Make sure you have Pipenv, Pipenv.lock & server.py in the folder.

4. Deploy our Lambda API

Now that everything is prepared, we can deploy our application to AWS using the CDK CLI.

$ cdk deploy

When this operation is complete you'll see a API Gateway URL displayed. This is your new API endpoint!!

5. Test our API

Now we need to test the functionality of your new API.

In our hypothetical example, when a train departs a station we call a POST operation to the /update endpoint:

$ http POST https://9vszbm10h9.execute-api.eu-west-1.amazonaws.com/prod/update
{
  "new_departure": "Edinburgh",
  "time": "18:46"
}

We can then retrieve the last departed train and the time that it departed using the /last_departed endpoint:

$ http GET https://9vszbm10h9.execute-api.eu-west-1.amazonaws.com/prod/last_departed
{
  "departure_time": "18:46",
  "last_departed": "Edinburgh"
}

That's it! You've now deployed a serverless Python API to API Gateway!

Clean Up

If you want to remove the resources for this sample. Use the AWS CDK as follows:

$ cdk destroy

kasukur/example-serverless-python-api