Defining Fargate containers with the AWS CDK

ara225 profile image Anna Aitchison ・1 min read


This is an attempt to break down the steps needed to define Fargate containers with the CDK.

Code examples are in Python, but should be easy enough to translate if needed. Fargate containers can be slightly more difficult to work with than most things in the CDK, as they require you to create an IAM policy separately. You also need to create a dockerfile separately. You can visit my GitHub project https://github.com/Ara225/serverless-audiobook-maker to see this code in situ

1.Import modules

from aws_cdk import (core

2.Create an IAM execution role.

role = aws_iam.Role(self, "FargateContainerRole", assumed_by=aws_iam.ServicePrincipal("ecs-tasks.amazonaws.com"))

3.Add policies granting the permissions you need

role.add_to_policy(aws_iam.PolicyStatement(actions=["s3:PutObject"], resources=[VideoUploadBucket.bucket_arn+"/*"]))

4.Create a VPC for your container to run in

vpc = aws_ec2.Vpc(self, "CdkFargateVpc", max_azs=2)

5.Create a cluster

cluster = aws_ecs.Cluster(self, 'FargateCluster', vpc=vpc)

6.Add a container image

image = aws_ecs.ContainerImage.from_asset(PATH_TO_FOLDER_CONTAINING_DOCKER_FILE)

7.Create a task definition

task_definition = aws_ecs.FargateTaskDefinition(
    self, "FargateContainerTaskDefinition", execution_role=role, task_role=role, cpu=1024, memory_limit_mib=3072

8.Add container to the task definition. Don't think logging is necessary but is useful

container = task_definition.add_container(
    "Container", image=image,

9.Optional - Add a port mapping

port_mapping = aws_ecs.PortMapping(container_port=80, host_port=80)

Posted on by:

ara225 profile

Anna Aitchison


Linux engineer by day, developer by night. Lover of all things serverless, cloud and Python.


markdown guide