Flask CRUD API on AWS

In this article, we’ll deploy the Flask application we created in the previous article on AWS, using:

Amazon RDS (Relational Database Service)
Amazon ECR (Elastic Container Registry)
Amazon ECS (Elastic Container Service)

Video Version:

Steps

Create RDS instance of Postgres
Test the RDS instance
Clone and build the TinyStacks GitHub repository
Create ECR repository
Tag the Docker image and push it to the ECR repository
Deploy to ECS using the ECR image
Test with Postman (and TablePlus)

Create RDS Instance

Go to the AWS Management Consoleonsole and search for RDS:

Click on Create Database.

From the list of available options, select:

Standard
Postgres version 12.5
Free Tier (If you’re a new customer or haven’t used RDS before, you can get up to 750 hours of usage free)

Choose a name for the instance (basically the name of the database), the master username (postgres), and the password (postgres). Confirm the password. (Make sure to use a more secure password in a production environment!)

You can leave the rest as it is, except for Connectivity:

We’ll set this up soon.

You can leave the rest as it is. Just double-check that the estimated monthly cost shows "Free Tier" (assuming you still have Free Tier credits). . Note that the cost of RDS instances may vary over time.

Then click Create Database:

While you wait, click on the database name and on the security group.

Click on the security group id, then on Edit Inbound rules. This will let us test our database with an external tool. If you want, you can skip this step. If you do, be sure that both the RDS instance and your app are in the same security group.

Now, add another inbound rule and make the RDS instance accessible:

Now we are ready to test the RDS instance.

Test the RDS instance

To test the RDS instance, we will use TablePlus, but you can use any tool you want. Alternative tools include:

pgadmin
psql

Create a new PostgreSQL connection:

Fill in all the information required to connect:

host: (the endpoint of the RDS instance)
port: 5432
username: postgres
password: postgres
database: postgres

You can disable SSL mode. (But make sure to leave SSL enabled when working in a production environment!)

Click Test Connection. If you see a connection is ok message, you’re good to go. The database is empty for now.

Clone the repository, change branch, and build the Docker image

Clone the aws-docker-templates-flask repository from Tinystacks:

git clone https://github.com/tinystacks/aws-docker-templates-flask.git

Step into the directory

cd aws-docker-templates-flask

Now, switch your GitHub branch. You can check all the branches with the command git branch -a.

git checkout flask-local-postgres

Then open this folder with your favorite IDE. IF you have Visual Studio code, you can type:

code .

Open a terminal and, in the folder where the docker-compose.yml file is located, type:

docker compose build

ECR

aws ecr get-login-password --region <REGION> | docker login --username AWS --password-stdin <AWS_ACCOUNT_ID>.dkr.ecr.<REGION>.amazonaws.com

Create a new ECR repository using the prompt. Use your <REGION>:

aws ecr create-repository --repository-name flask-app --region <REGION>

Then you can check the Amazon Console registries on ECR:

If you click on the repository name, you won’t see any images. That’s what we expect (for now).

Tag and push the Docker image on ECR

You can use the value <IMAGE_TAG> to create an image with the tag latest.

docker tag pythonapp <IMAGE_TAG>

Now push the image to the ECR repository:

docker push <IMAGE_TAG>

And check the image on the ECR repository:

ECS

Now, let's deploy our service using Amazon Elastic Container Service (ECS).

Search for ecs on the AWS Management Console:

Click on Get Started:

Click on Custom and then click Configure:

This will let us configure the container task we want to run in our cluster.

Choose:

Container name: any you want
Image: the <IMAGE_TAG> (latest)
Port: 80

Next, in Advanced Configuration add an environment variable as defined in the docker-compose.yml file:

For the values here, use:

Key: <DATABASE_URL>
Value: postgresql://postgres:postgres@<RDS_INSTANCE_IP>:5432/postgres

Remember to replace the database entry with the <RDS_INSTANCE_IP> for your instance. You can find your <RDS_INSTANCE_IP> on TablePlus or RDS.

Leave the rest as it is and click Update.

On Task definition, click Next:

Likewise, on Service, click Next:

On Cluster configuration, define a name for your cluster and click Next:

You’re all set! !Click Create:

Creation will take a couple of minutes. Once all the check marks have turned green, click View Service:

Click on the task and copy its Public IP. We will need that to test our application.

Test With Postman and Tableplus

First, let’s get all the items. If we get an empty list [], it means that the table has been created and we are getting no items. That’s what we expect, since we just created the table with no data.

GET Request at the endpoint: /items

To create a new item, make a POST request at the endpoint /items:

Let's create 2 more items:

To get a single item, make a GET request, appending the id of the item we want to retrieve at the end of the /items path. For example /items/2 retrieves the item with id = 2.
(Don't mind the body in this request, that was from the last post request.)