[PYTHON] Building an AWS Fargate Service with AWS CDK

Introduction

I would like to build an AWS Fargate Service using AWS CDK (Python).

The code created this time is here

↑ Create the following configuration with this code.

Diagram

Preparation

--AWS account --IAM User for CDK

flow

1. Environment variables
2. cdk install
3. cdk init
4. Implementation
5. cdk deploy
6. [Operation check](# Operation check)
7. cdk destroy

cdk install

$ npm install -g aws-cdk

If cdk cannot be executed, check if it is in the PATH.

$ which node
/path/to/bin/node
$ export
...
PATH=/path/to/hoge:/path/to/fuga
...
$ export PATH=$PATH:/path/to/current/bin
$ export
...
PATH=/path/to/hoge:/path/to/fuga:/path/to/current/bin
...

Environment variable settings

export AWS_ACCESS_KEY_ID=hoge
export AWS_SECRET_ACCESS_KEY=fuga
export ROLE_ARN=arn:aws:iam::xxxxxxxxxxxx:role/ecsTaskExecutionRole
export ECR_REGISOTRY=xxxxxxxxxxxx:role.dkr.ecr.ap-northeast-1.amazonaws.com/django-app:latest

cdk init

$ mkdir /path/to/cdk_fargate
$ cd /path/to/cdk_fargate
$ cdk init --language python

Implementation

$ pwd
/path/to/cdk_fargate
$ cd cdk_fargate/
vi cdk_fargate_stack.py  

cdk_fargate_stack.py is here

VPC creation

--Create a VPC by specifying two Availability Zones.

Create ECS Cluster

--Create an ECS Cluster by specifying the created VPC.

Fargate Service creation

-Build a Fargate Service using ecs_pattern.ApplicationLoadBalancedFargateService. --This time, see the image of Nginx.

Output --Output ALB DNS

import os
from dotenv import load_dotenv

from aws_cdk import (
    aws_ec2 as ec2,
    aws_ecs as ecs,
    aws_iam as iam,
    aws_ecs_patterns as ecs_patterns,
    core,
)

load_dotenv()

ROLE_ARN = os.environ["ROLE_ARN"]
ECR_REGISOTRY = os.environ["ECR_REGISOTRY"]


class CdkFargateStack(core.Stack):
    def __init__(self, scope: core.Construct, id: str, **kwargs) -> None:
        super().__init__(scope, id, **kwargs)

        vpc = ec2.Vpc(self, "CdkFargateVpc", max_azs=2)

        cluster = ecs.Cluster(self, "Ec2Cluster", vpc=vpc)

        role = iam.Role.from_role_arn(self, "Role", ROLE_ARN)
        image = ecs.ContainerImage.from_registry(ECR_REGISOTRY)
        task_definition = ecs.FargateTaskDefinition(
            scope=self, id="TaskDefinition", execution_role=role, task_role=role
        )
        port_mapping = ecs.PortMapping(container_port=80, host_port=80)
        container = task_definition.add_container(
            id="Container", image=image
        ).add_port_mappings(port_mapping)

        fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(
            self, "FargateService", cluster=cluster, task_definition=task_definition
        )

        core.CfnOutput(
            self,
            "LoadBalancerDNS",
            value=fargate_service.load_balancer.load_balancer_dns_name,
        )

cdk deploy

$ export AWS_ACCESS_KEY_ID=hoge
$ export AWS_SECRET_ACCESS_KEY=fuga
$ cdk deploy

Operation check

$ curl cdk-f-farga-xxxxxxxxx.ap-northeast-1.elb.amazonaws.com

cdk destroy

$ cdk destory

All resources have been deleted!

What I want to do next

--Image management using ECR --Custom domain support using ACM and Route53

Summary

The team develops using Python, but I found it wonderful to be able to manage the configuration using a common language. I would like to continue to promote configuration management using CDK. I couldn't even implement custom domain settings

References

• https://docs.aws.amazon.com/ja_jp/cdk/latest/guide/ecs_example.html
• https://docs.aws.amazon.com/cdk/api/latest/docs/aws-ecs-readme.html
• https://docs.aws.amazon.com/cdk/api/latest/python/aws_cdk.aws_ecs_patterns.html
• https://github.com/aws-samples/aws-cdk-examples