ECS, ALB and API Gateway

Create ECS Cluster and Application Load Balancer

1. Access VSCode
   • If you are in the cfn-lint virtual environment, exit using deactivate
   • Go back to the home directory /home/ec2-user
   • Create a new folder named cdk-workshop-02 and enter that folder
   • Create a new CDK project with python language

cd ~
mkdir cdk-workshop-02
cd cdk-workshop-02
cdk init app --language python

2. Open the newly created CDK file cdk-workshop-02/cdk_workshop_02_stack.py and add these libraries

from aws_cdk import (
  Stack,
  aws_ecs as ecs,
  aws_ecs_patterns as ecsp,
  aws_apigateway as apigateway
)
from constructs import Construct

3. In cdk-workshop-02/cdk_workshop_02_stack.py, declare an ECS cluster with an Application Load Balancer (ALB) inside the __init__() function

      # Declare a Load Balancer Fargate 
      lb_fargate_service = ecsp.ApplicationLoadBalancedFargateService(
				self, 
				"Workshop02-MyWebServer",
        task_image_options=ecsp.ApplicationLoadBalancedTaskImageOptions(
            image=ecs.ContainerImage.from_registry("nginxdemos/hello")
        ),
        public_load_balancer=True,
        desired_count=3
      )

ecsp.ApplicationLoadBalancedFargateService will create the following resources:

• A new VPC with default configuration (subnet, Internet Gateway, RouteTable, EIP, Nat Gateway, …)
• A new ECS cluster created in that VPC, running on Fargate
• An ECS Task defined with the nginxdemos/hello image from Dockerhub
• An ECS Container group initialized from the ECS Task, with a desired count of 3
• An Application Load Balancer placed in front of that ECS cluster
• The necessary roles and security groups

Configure API Gateway

4. Declare an API Gateway

# Define API Gateway
api = apigateway.RestApi(self, "ProxyToLBECS")

5. Configure the API Gateway resource to point to the DNS of the Application Load Balancer

# Get the DNS value of the Application Load Balancer 
lb = lb_fargate_service.load_balancer
lb_dns = lb_fargate_service.load_balancer.load_balancer_dns_name

# Add resource and method for proxy request
proxy = api.root.add_resource("ecs")
proxy.add_method("GET", apigateway.HttpIntegration(f"http://{lb_dns}"))

You can read more about the methods and properties of ApplicationLoadBalancedFargateService at the official CDK documentation of AWS (Python)

6. Review the code in cdk-workshop-02/cdk_workshop_02_stack.py

from aws_cdk import (
    Stack,
    aws_ecs as ecs,
    aws_ecs_patterns as ecsp,
    aws_apigateway as apigateway
)
from constructs import Construct

class CdkWorkshop02Stack(Stack):

    def __init__(self, scope: Construct, construct_id: str, **kwargs) -> None:
        super().__init__(scope, construct_id, **kwargs)
        
        # Declare a Load Balancer Fargate 
        lb_fargate_service = ecsp.ApplicationLoadBalancedFargateService(
          self, 
          "Workshop02-MyWebServer",
          task_image_options=ecsp.ApplicationLoadBalancedTaskImageOptions(
              image=ecs.ContainerImage.from_registry("nginxdemos/hello")),
          public_load_balancer=True,
          desired_count=3
        )
        
        # Define API Gateway
        api = apigateway.RestApi(self, "ProxyToLBECS")
        
        # Get the DNS value of the Application Load Balancer 
        lb = lb_fargate_service.load_balancer
        lb_dns = lb_fargate_service.load_balancer.load_balancer_dns_name
        
        # Add resource and method for proxy request
        proxy = api.root.add_resource("ecs")
        proxy.add_method("GET", apigateway.HttpIntegration(f"http://{lb_dns}"))

Initialize a Python Virtual Environment

Before we can deploy to the Cloud, we need to:

• Initialize a virtual environment in the cdk-workshop-02 folder. The server has both the Amazon Linux 2023 system Python and the self-installed Python 3.14, so we isolate the project’s libraries in a venv to avoid conflicts.
• Install the packages listed in requirements.txt

7. Create a virtual environment using Python 3.14 right inside the cdk-workshop-02 folder.

python3.14 -m venv venv
source venv/bin/activate

After activating, verify that the venv is correctly using Python 3.14:

python --version

8. Install the required libraries from requirements.txt. Use python -m pip to ensure they are installed in the correct venv:

python -m pip install -r requirements.txt

Deploy

cd ~/cdk-workshop-02
source venv/bin/activate

9. Check what we have created before deploying

cdk diff

10. Before deploying, we need to bootstrap the EC2 environment

cdk bootstrap

11. Then, deploy the declared stack

cdk deploy

Press y to continue

12. Finish and check the results

Access the API Gateway link (ProxyToLBECSEndpoint), and append the path /ecs. You will be able to access the web server deployed in ECS.

To verify the Application Load Balancer is working, refresh the browser a few times, and you will see the destination of the web page change. In the previous section, we set the parameter desired_count=3, so there are 3 containers deployed in parallel behind the ALB. You will see the Server address change across browser refreshes.

If you have reached this point, congratulations on completing the first part of the CDK Basic 2 lab. In the next section, we will use CDK to deploy a Lambda function behind an API Gateway resource.