Building Serverless Python Apps Using AWS Chalice

Why Chalice?

For Python developers accustomed to the Flask web framework, Chalice should be a breeze in terms of building and shipping your first app. Highly inspired by Flask, Chalice keeps it pretty minimalist in terms of defining what the service should be like and finally making an executable package of the same.

Enough theory! Let’s start with a basic hello-world app and kick-start our serverless journey.

Project Setup

Before diving into Chalice, you’ll set up a working environment on your local machine, which will set you up for the rest of the tutorial.

First, create and activate a virtual environment and install Chalice:

$ python3.6 -m venv env
$ source env/bin/activate
(env)$ pip install chalice

Follow our comprehensive guide on the Pipenv packaging tool.

Now that you have Chalice installed on your virtual environment, let’s use the Chalice CLI to generate some boilerplate code:

(env)$ chalice new-project

Enter the name of the project when prompted and hit return. A new directory is created with that name:

<project-name>/
|
├── .chalice/
│   └── config.json
|
├── .gitignore
├── app.py
└── requirements.txt

See how minimalist the Chalice codebase is. A .chalice directory, app.py, and requirements.txt is all that it requires to have a serverless app up and running. Let’s quickly run the app on our local machine.

Chalice CLI consists of really great utility functions allowing you to perform a number of operations from running locally to deploying in a Lambda environment.

Build and Run Locally

You can simulate the app by running it locally using the local utility of Chalice:

(env)$ chalice local
Serving on 127.0.0.1:8000

By default, Chalice runs on port 8000. We can now check the index route by making a curl request to http://localhost:8000/:

$ curl -X GET http://localhost:8000/
{"hello": "world"}

Now if we look at app.py, we can appreciate the simplicity with which Chalice allows you to build a serverless service. All the complex stuff is handled by the decorators:

from chalice import Chalice
app = Chalice(app_name='serverless-sms-service')

@app.route('/')
def index():
    return {'hello': 'world'}

Now, let’s move on to deploying our app on the AWS Lambda.

Deploy on AWS Lambda

Chalice makes deploying your serverless app completely effortless. Using the deploy utility, you can simply instruct Chalice to deploy and create a Lambda function that can be accessible via a REST API.

Before we begin deployment, we need to make sure we have our AWS credentials in place, usually located at ~/.aws/config. The contents of the file look as follows:

[default]
aws_access_key_id=<your-access-key-id>
aws_secret_access_key=<your-secret-access-key>
region=<your-region>

With AWS credentials in place, let’s begin our deployment process with just a single command:

(env)$ chalice deploy
Creating deployment package.
Updating policy for IAM role: hello-world-dev
Creating lambda function: hello-world-dev
Creating Rest API
Resources deployed:
  - Lambda ARN: arn:aws:lambda:ap-south-1:679337104153:function:hello-world-dev
  - Rest API URL: https://fqcdyzvytc.execute-api.ap-south-1.amazonaws.com/api/

Wow! Yes, it is really this easy to get your serverless app up and running. To verify simply make a curl request on the generated Rest API URL:

$ curl -X GET https://fqcdyzvytc.execute-api.ap-south-1.amazonaws.com/api/
{"hello": "world"}

Typically, this is all that you need to get your serverless app up and running. You can also go to your AWS console and see the Lambda function created under the Lambda service section. Each Lambda service has a unique REST API endpoint that can be consumed in any web application.

Next, you will begin building your Serverless SMS Sender service using Twilio as an SMS service provider.

Build and Run Locally

Now you can run your app on your machine using the local utility and verify that everything is working fine:

(env)$ chalice local

Now make a curl POST request to http://localhost:8000/service/sms/send with a specific payload and test the app locally:

$ curl -H "Content-Type: application/json" -X POST -d '{"msg": "hey mate!!!"}' http://localhost:8000/service/sms/send

The above request responds as follows:

{
  "status": "success",
  "data": "SM60f11033de4f4e39b1c193025bcd5cd8",
  "message": "SMS successfully sent"
}

The response indicates that the message was successfully sent. Now, let’s move on to deploying the app on AWS Lambda.

Deploy on AWS Lambda

As suggested in the previous deployment section, you just need to issue the following command:

(env)$ chalice deploy
Creating deployment package.
Updating policy for IAM role: sms-shooter-dev
Creating lambda function: sms-shooter-dev
Creating Rest API
Resources deployed:
  - Lambda ARN: arn:aws:lambda:ap-south-1:679337104153:function:sms-shooter-dev
  - Rest API URL: https://qtvndnjdyc.execute-api.ap-south-1.amazonaws.com/api/

twilio==6.18.1

Once all the package dependencies are sorted, you need to make sure all the environment variables are also shipped along and set correctly during the Lambda runtime. To do so, you have to add all the environment variables in .chalice/config.json in the following manner:

{
  "version": "2.0",
  "app_name": "sms-shooter",
  "stages": {
    "dev": {
      "api_gateway_stage": "api",
      "environment_variables": {
          "ACCOUNT_SID": "<your-account-sid>",
          "AUTH_TOKEN": "<your-auth-token>",
          "FROM_NUMBER": "<source-number>",
          "TO_NUMBER": "<destination-number>"
      }
    }
  }
}

Now we’re good to deploy:

Creating deployment package.
Updating policy for IAM role: sms-shooter-dev
Updating lambda function: sms-shooter-dev
Updating rest API
Resources deployed:
  - Lambda ARN: arn:aws:lambda:ap-south-1:679337104153:function:sms-shooter-dev
  - Rest API URL: https://fqcdyzvytc.execute-api.ap-south-1.amazonaws.com/api/

Do a sanity check by making a curl request to the generated API endpoint:

$ curl -H "Content-Type: application/json" -X POST -d '{"msg": "hey mate!!!"}' https://fqcdyzvytc.execute-api.ap-south-1.amazonaws.com/api/service/sms/send

The above request responds as expected:

{
  "status": "success",
  "data": "SM60f11033de4f4e39b1c193025bcd5cd8",
  "message": "SMS successfully sent"
}

Now, you have a completely serverless SMS sending service up and running. With the front end of this service being a REST API, it can be used in other applications as a plug-and-play feature that is scalable, secure, and reliable.

Build and Run Locally

Run the app once again using the local utility:

(env)$ chalice local

Make a curl request and verify. Once that’s done, move on to deployment.

Deploy on AWS Lambda

Once you are sure everything is working as expected, you can now finally deploy your app:

(env)$ chalice deploy

As usual, the command executes successfully and you can verify the endpoint.