CDK Construct library for higher-level ECS Constructs

This library provides higher-level Amazon ECS constructs which follow common architectural patterns. It contains:

• Application Load Balanced Services

• Network Load Balanced Services

• Queue Processing Services

• Scheduled Tasks (cron jobs)

• Additional Examples

Application Load Balanced Services

To define an Amazon ECS service that is behind an application load balancer, instantiate one of the following:

• ApplicationLoadBalancedEc2Service

# cluster: ecs.Cluster

load_balanced_ecs_service = ecs_patterns.ApplicationLoadBalancedEc2Service(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("test"),
        environment={
            "TEST_ENVIRONMENT_VARIABLE1": "test environment variable 1 value",
            "TEST_ENVIRONMENT_VARIABLE2": "test environment variable 2 value"
        },
        command=["command"],
        entry_point=["entry", "point"]
    ),
    desired_count=2
)

• ApplicationLoadBalancedFargateService

# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        command=["command"],
        entry_point=["entry", "point"]
    )
)

load_balanced_fargate_service.target_group.configure_health_check(
    path="/custom-health-path"
)

Instead of providing a cluster you can specify a VPC and CDK will create a new ECS cluster. If you deploy multiple services CDK will only create one cluster per VPC.

You can omit cluster and vpc to let CDK create a new VPC with two AZs and create a cluster inside this VPC.

You can customize the health check for your target group; otherwise it defaults to HTTP over port 80 hitting path /.

You can customize the health check configuration of the container via the healthCheck property; otherwise it defaults to the health check configuration from the container.

Fargate services will use the LATEST platform version by default, but you can override by providing a value for the platformVersion property in the constructor.

Fargate services use the default VPC Security Group unless one or more are provided using the securityGroups property in the constructor.

By setting redirectHTTP to true, CDK will automatically create a listener on port 80 that redirects HTTP traffic to the HTTPS port.

If you specify the option recordType you can decide if you want the construct to use CNAME or Route53-Aliases as record sets.

If you need to encrypt the traffic between the load balancer and the ECS tasks, you can set the targetProtocol to HTTPS.

Additionally, if more than one application target group are needed, instantiate one of the following:

• ApplicationMultipleTargetGroupsEc2Service

# One application load balancer with one listener and two target groups.
# cluster: ecs.Cluster

load_balanced_ec2_service = ecs_patterns.ApplicationMultipleTargetGroupsEc2Service(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=256,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageProps(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    target_groups=[ecsPatterns.ApplicationTargetProps(
        container_port=80
    ), ecsPatterns.ApplicationTargetProps(
        container_port=90,
        path_pattern="a/b/c",
        priority=10
    )
    ]
)

• ApplicationMultipleTargetGroupsFargateService

# One application load balancer with one listener and two target groups.
# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.ApplicationMultipleTargetGroupsFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageProps(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    target_groups=[ecsPatterns.ApplicationTargetProps(
        container_port=80
    ), ecsPatterns.ApplicationTargetProps(
        container_port=90,
        path_pattern="a/b/c",
        priority=10
    )
    ]
)

Network Load Balanced Services

To define an Amazon ECS service that is behind a network load balancer, instantiate one of the following:

• NetworkLoadBalancedEc2Service

# cluster: ecs.Cluster

load_balanced_ecs_service = ecs_patterns.NetworkLoadBalancedEc2Service(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("test"),
        environment={
            "TEST_ENVIRONMENT_VARIABLE1": "test environment variable 1 value",
            "TEST_ENVIRONMENT_VARIABLE2": "test environment variable 2 value"
        }
    ),
    desired_count=2
)

• NetworkLoadBalancedFargateService

# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.NetworkLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    cpu=512,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    )
)

The CDK will create a new Amazon ECS cluster if you specify a VPC and omit cluster. If you deploy multiple services the CDK will only create one cluster per VPC.

If cluster and vpc are omitted, the CDK creates a new VPC with subnets in two Availability Zones and a cluster within this VPC.

If you specify the option recordType you can decide if you want the construct to use CNAME or Route53-Aliases as record sets.

Additionally, if more than one network target group is needed, instantiate one of the following:

• NetworkMultipleTargetGroupsEc2Service

# Two network load balancers, each with their own listener and target group.
# cluster: ecs.Cluster

load_balanced_ec2_service = ecs_patterns.NetworkMultipleTargetGroupsEc2Service(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=256,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageProps(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    load_balancers=[ecsPatterns.NetworkLoadBalancerProps(
        name="lb1",
        listeners=[ecsPatterns.NetworkListenerProps(
            name="listener1"
        )
        ]
    ), ecsPatterns.NetworkLoadBalancerProps(
        name="lb2",
        listeners=[ecsPatterns.NetworkListenerProps(
            name="listener2"
        )
        ]
    )
    ],
    target_groups=[ecsPatterns.NetworkTargetProps(
        container_port=80,
        listener="listener1"
    ), ecsPatterns.NetworkTargetProps(
        container_port=90,
        listener="listener2"
    )
    ]
)

• NetworkMultipleTargetGroupsFargateService

# Two network load balancers, each with their own listener and target group.
# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.NetworkMultipleTargetGroupsFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=512,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageProps(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    load_balancers=[ecsPatterns.NetworkLoadBalancerProps(
        name="lb1",
        listeners=[ecsPatterns.NetworkListenerProps(
            name="listener1"
        )
        ]
    ), ecsPatterns.NetworkLoadBalancerProps(
        name="lb2",
        listeners=[ecsPatterns.NetworkListenerProps(
            name="listener2"
        )
        ]
    )
    ],
    target_groups=[ecsPatterns.NetworkTargetProps(
        container_port=80,
        listener="listener1"
    ), ecsPatterns.NetworkTargetProps(
        container_port=90,
        listener="listener2"
    )
    ]
)

Queue Processing Services

To define a service that creates a queue and reads from that queue, instantiate one of the following:

• QueueProcessingEc2Service

# cluster: ecs.Cluster

queue_processing_ec2_service = ecs_patterns.QueueProcessingEc2Service(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    image=ecs.ContainerImage.from_registry("test"),
    command=["-c", "4", "amazon.com"],
    enable_logging=False,
    desired_task_count=2,
    environment={
        "TEST_ENVIRONMENT_VARIABLE1": "test environment variable 1 value",
        "TEST_ENVIRONMENT_VARIABLE2": "test environment variable 2 value"
    },
    max_scaling_capacity=5,
    container_name="test"
)

• QueueProcessingFargateService

# cluster: ecs.Cluster

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    command=["-c", "4", "amazon.com"],
    enable_logging=False,
    desired_task_count=2,
    environment={
        "TEST_ENVIRONMENT_VARIABLE1": "test environment variable 1 value",
        "TEST_ENVIRONMENT_VARIABLE2": "test environment variable 2 value"
    },
    max_scaling_capacity=5,
    container_name="test"
)

when queue not provided by user, CDK will create a primary queue and a dead letter queue with default redrive policy and attach permission to the task to be able to access the primary queue.

NOTE: QueueProcessingFargateService adds a CPU Based scaling strategy by default. You can turn this off by setting disableCpuBasedScaling: true.

# cluster: ecs.Cluster

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    command=["-c", "4", "amazon.com"],
    enable_logging=False,
    desired_task_count=2,
    environment={
        "TEST_ENVIRONMENT_VARIABLE1": "test environment variable 1 value",
        "TEST_ENVIRONMENT_VARIABLE2": "test environment variable 2 value"
    },
    max_scaling_capacity=5,
    container_name="test",
    disable_cpu_based_scaling=True
)

To specify a custom target CPU utilization percentage for the scaling strategy use the cpuTargetUtilizationPercent property:

# cluster: ecs.Cluster

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    command=["-c", "4", "amazon.com"],
    enable_logging=False,
    desired_task_count=2,
    environment={},
    max_scaling_capacity=5,
    container_name="test",
    cpu_target_utilization_percent=90
)

Scheduled Tasks

To define a task that runs periodically, there are 2 options:

• ScheduledEc2Task

# Instantiate an Amazon EC2 Task to run at a scheduled interval
# cluster: ecs.Cluster

ecs_scheduled_task = ecs_patterns.ScheduledEc2Task(self, "ScheduledTask",
    cluster=cluster,
    scheduled_ec2_task_image_options=ecsPatterns.ScheduledEc2TaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        memory_limit_mi_b=256,
        environment={"name": "TRIGGER", "value": "CloudWatch Events"}
    ),
    schedule=appscaling.Schedule.expression("rate(1 minute)"),
    enabled=True,
    rule_name="sample-scheduled-task-rule"
)

• ScheduledFargateTask

# cluster: ecs.Cluster

scheduled_fargate_task = ecs_patterns.ScheduledFargateTask(self, "ScheduledFargateTask",
    cluster=cluster,
    scheduled_fargate_task_image_options=ecsPatterns.ScheduledFargateTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        memory_limit_mi_b=512
    ),
    schedule=appscaling.Schedule.expression("rate(1 minute)"),
    platform_version=ecs.FargatePlatformVersion.LATEST
)

Additional Examples

In addition to using the constructs, users can also add logic to customize these constructs:

Configure HTTPS on an ApplicationLoadBalancedFargateService

from aws_cdk.aws_route53 import HostedZone
from aws_cdk.aws_certificatemanager import Certificate
from aws_cdk.aws_elasticloadbalancingv2 import SslPolicy

# vpc: ec2.Vpc
# cluster: ecs.Cluster


domain_zone = HostedZone.from_lookup(self, "Zone", domain_name="example.com")
certificate = Certificate.from_certificate_arn(self, "Cert", "arn:aws:acm:us-east-1:123456:certificate/abcdefg")
load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    vpc=vpc,
    cluster=cluster,
    certificate=certificate,
    ssl_policy=SslPolicy.RECOMMENDED,
    domain_name="api.example.com",
    domain_zone=domain_zone,
    redirect_hTTP=True,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    )
)

Set capacityProviderStrategies for ApplicationLoadBalancedFargateService

# cluster: ecs.Cluster

cluster.enable_fargate_capacity_providers()

load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    capacity_provider_strategies=[ecs.CapacityProviderStrategy(
        capacity_provider="FARGATE_SPOT",
        weight=2,
        base=0
    ), ecs.CapacityProviderStrategy(
        capacity_provider="FARGATE",
        weight=1,
        base=1
    )
    ]
)

Add Schedule-Based Auto-Scaling to an ApplicationLoadBalancedFargateService

# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    desired_count=1,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    )
)

scalable_target = load_balanced_fargate_service.service.auto_scale_task_count(
    min_capacity=5,
    max_capacity=20
)

scalable_target.scale_on_schedule("DaytimeScaleDown",
    schedule=appscaling.Schedule.cron(hour="8", minute="0"),
    min_capacity=1
)

scalable_target.scale_on_schedule("EveningRushScaleUp",
    schedule=appscaling.Schedule.cron(hour="20", minute="0"),
    min_capacity=10
)

Add Metric-Based Auto-Scaling to an ApplicationLoadBalancedFargateService

# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    desired_count=1,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    )
)

scalable_target = load_balanced_fargate_service.service.auto_scale_task_count(
    min_capacity=1,
    max_capacity=20
)

scalable_target.scale_on_cpu_utilization("CpuScaling",
    target_utilization_percent=50
)

scalable_target.scale_on_memory_utilization("MemoryScaling",
    target_utilization_percent=50
)

Change the default Deployment Controller

# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    desired_count=1,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    deployment_controller=ecs.DeploymentController(
        type=ecs.DeploymentControllerType.CODE_DEPLOY
    )
)

Deployment circuit breaker and rollback

Amazon ECS deployment circuit breaker automatically rolls back unhealthy service deployments without the need for manual intervention. Use circuitBreaker to enable deployment circuit breaker and optionally enable rollback for automatic rollback. See Using the deployment circuit breaker for more details.

# cluster: ecs.Cluster

service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    desired_count=1,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    circuit_breaker=ecs.DeploymentCircuitBreaker(rollback=True)
)

Set deployment configuration on QueueProcessingService

# cluster: ecs.Cluster

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    command=["-c", "4", "amazon.com"],
    enable_logging=False,
    desired_task_count=2,
    environment={},
    max_scaling_capacity=5,
    max_healthy_percent=200,
    min_healthy_percent=66
)

Set taskSubnets and securityGroups for QueueProcessingFargateService

# vpc: ec2.Vpc
# security_group: ec2.SecurityGroup

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    vpc=vpc,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    security_groups=[security_group],
    task_subnets=ec2.SubnetSelection(subnet_type=ec2.SubnetType.PRIVATE_ISOLATED)
)

Define tasks with public IPs for QueueProcessingFargateService

# vpc: ec2.Vpc

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    vpc=vpc,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    assign_public_ip=True
)

Define tasks with custom queue parameters for QueueProcessingFargateService

# vpc: ec2.Vpc

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    vpc=vpc,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    max_receive_count=42,
    retention_period=Duration.days(7),
    visibility_timeout=Duration.minutes(5)
)

Set cooldown for QueueProcessingFargateService

The cooldown period is the amount of time to wait for a previous scaling activity to take effect. To specify something other than the default cooldown period of 300 seconds, use the cooldown parameter:

# vpc: ec2.Vpc

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    vpc=vpc,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    assign_public_ip=True,
    cooldown=Duration.seconds(500)
)

Set capacityProviderStrategies for QueueProcessingFargateService

# cluster: ecs.Cluster

cluster.enable_fargate_capacity_providers()

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    capacity_provider_strategies=[ecs.CapacityProviderStrategy(
        capacity_provider="FARGATE_SPOT",
        weight=2
    ), ecs.CapacityProviderStrategy(
        capacity_provider="FARGATE",
        weight=1
    )
    ]
)

Set a custom container-level Healthcheck for QueueProcessingFargateService

# vpc: ec2.Vpc
# security_group: ec2.SecurityGroup

queue_processing_fargate_service = ecs_patterns.QueueProcessingFargateService(self, "Service",
    vpc=vpc,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    health_check=ecs.HealthCheck(
        command=["CMD-SHELL", "curl -f http://localhost/ || exit 1"],
        # the properties below are optional
        interval=Duration.minutes(30),
        retries=123,
        start_period=Duration.minutes(30),
        timeout=Duration.minutes(30)
    )
)

Set capacityProviderStrategies for QueueProcessingEc2Service

import aws_cdk.aws_autoscaling as autoscaling


vpc = ec2.Vpc(self, "Vpc", max_azs=1)
cluster = ecs.Cluster(self, "EcsCluster", vpc=vpc)
auto_scaling_group = autoscaling.AutoScalingGroup(self, "asg",
    vpc=vpc,
    instance_type=ec2.InstanceType.of(ec2.InstanceClass.BURSTABLE2, ec2.InstanceSize.MICRO),
    machine_image=ecs.EcsOptimizedImage.amazon_linux2()
)
capacity_provider = ecs.AsgCapacityProvider(self, "provider",
    auto_scaling_group=auto_scaling_group
)
cluster.add_asg_capacity_provider(capacity_provider)

queue_processing_ec2_service = ecs_patterns.QueueProcessingEc2Service(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=512,
    image=ecs.ContainerImage.from_registry("test"),
    capacity_provider_strategies=[ecs.CapacityProviderStrategy(
        capacity_provider=capacity_provider.capacity_provider_name
    )
    ]
)

Select specific vpc subnets for ApplicationLoadBalancedFargateService

# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    desired_count=1,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    task_subnets=ec2.SubnetSelection(
        subnets=[ec2.Subnet.from_subnet_id(self, "subnet", "VpcISOLATEDSubnet1Subnet80F07FA0")]
    )
)

Select idleTimeout for ApplicationLoadBalancedFargateService

# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    desired_count=1,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    idle_timeout=Duration.seconds(120)
)

Select idleTimeout for ApplicationMultipleTargetGroupsFargateService

from aws_cdk.aws_certificatemanager import Certificate
from aws_cdk.aws_ec2 import InstanceType
from aws_cdk.aws_ecs import Cluster, ContainerImage
from aws_cdk.aws_elasticloadbalancingv2 import ApplicationProtocol, SslPolicy
from aws_cdk.aws_route53 import PublicHostedZone

vpc = ec2.Vpc(self, "Vpc", max_azs=1)
load_balanced_fargate_service = ecs_patterns.ApplicationMultipleTargetGroupsFargateService(self, "myService",
    cluster=ecs.Cluster(self, "EcsCluster", vpc=vpc),
    memory_limit_mi_b=256,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageProps(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    enable_execute_command=True,
    load_balancers=[ecsPatterns.ApplicationLoadBalancerProps(
        name="lb",
        idle_timeout=Duration.seconds(400),
        domain_name="api.example.com",
        domain_zone=PublicHostedZone(self, "HostedZone", zone_name="example.com"),
        listeners=[ecsPatterns.ApplicationListenerProps(
            name="listener",
            protocol=ApplicationProtocol.HTTPS,
            certificate=Certificate.from_certificate_arn(self, "Cert", "helloworld"),
            ssl_policy=SslPolicy.TLS12_EXT
        )
        ]
    ), ecsPatterns.ApplicationLoadBalancerProps(
        name="lb2",
        idle_timeout=Duration.seconds(120),
        domain_name="frontend.com",
        domain_zone=PublicHostedZone(self, "HostedZone", zone_name="frontend.com"),
        listeners=[ecsPatterns.ApplicationListenerProps(
            name="listener2",
            protocol=ApplicationProtocol.HTTPS,
            certificate=Certificate.from_certificate_arn(self, "Cert2", "helloworld"),
            ssl_policy=SslPolicy.TLS12_EXT
        )
        ]
    )
    ],
    target_groups=[ecsPatterns.ApplicationTargetProps(
        container_port=80,
        listener="listener"
    ), ecsPatterns.ApplicationTargetProps(
        container_port=90,
        path_pattern="a/b/c",
        priority=10,
        listener="listener"
    ), ecsPatterns.ApplicationTargetProps(
        container_port=443,
        listener="listener2"
    ), ecsPatterns.ApplicationTargetProps(
        container_port=80,
        path_pattern="a/b/c",
        priority=10,
        listener="listener2"
    )
    ]
)

Set health checks for ApplicationMultipleTargetGroupsFargateService

from aws_cdk.aws_certificatemanager import Certificate
from aws_cdk.aws_ec2 import InstanceType
from aws_cdk.aws_ecs import Cluster, ContainerImage
from aws_cdk.aws_elasticloadbalancingv2 import ApplicationProtocol, Protocol, SslPolicy
from aws_cdk.aws_route53 import PublicHostedZone

vpc = ec2.Vpc(self, "Vpc", max_azs=1)

load_balanced_fargate_service = ecs_patterns.ApplicationMultipleTargetGroupsFargateService(self, "myService",
    cluster=ecs.Cluster(self, "EcsCluster", vpc=vpc),
    memory_limit_mi_b=256,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageProps(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    enable_execute_command=True,
    load_balancers=[ecsPatterns.ApplicationLoadBalancerProps(
        name="lb",
        idle_timeout=Duration.seconds(400),
        domain_name="api.example.com",
        domain_zone=PublicHostedZone(self, "HostedZone", zone_name="example.com"),
        listeners=[ecsPatterns.ApplicationListenerProps(
            name="listener",
            protocol=ApplicationProtocol.HTTPS,
            certificate=Certificate.from_certificate_arn(self, "Cert", "helloworld"),
            ssl_policy=SslPolicy.TLS12_EXT
        )
        ]
    ), ecsPatterns.ApplicationLoadBalancerProps(
        name="lb2",
        idle_timeout=Duration.seconds(120),
        domain_name="frontend.com",
        domain_zone=PublicHostedZone(self, "HostedZone", zone_name="frontend.com"),
        listeners=[ecsPatterns.ApplicationListenerProps(
            name="listener2",
            protocol=ApplicationProtocol.HTTPS,
            certificate=Certificate.from_certificate_arn(self, "Cert2", "helloworld"),
            ssl_policy=SslPolicy.TLS12_EXT
        )
        ]
    )
    ],
    target_groups=[ecsPatterns.ApplicationTargetProps(
        container_port=80,
        listener="listener"
    ), ecsPatterns.ApplicationTargetProps(
        container_port=90,
        path_pattern="a/b/c",
        priority=10,
        listener="listener"
    ), ecsPatterns.ApplicationTargetProps(
        container_port=443,
        listener="listener2"
    ), ecsPatterns.ApplicationTargetProps(
        container_port=80,
        path_pattern="a/b/c",
        priority=10,
        listener="listener2"
    )
    ]
)

load_balanced_fargate_service.target_groups[0].configure_health_check(
    port="8050",
    protocol=Protocol.HTTP,
    healthy_threshold_count=2,
    unhealthy_threshold_count=2,
    timeout=Duration.seconds(10),
    interval=Duration.seconds(30),
    healthy_http_codes="200"
)

load_balanced_fargate_service.target_groups[1].configure_health_check(
    port="8050",
    protocol=Protocol.HTTP,
    healthy_threshold_count=2,
    unhealthy_threshold_count=2,
    timeout=Duration.seconds(10),
    interval=Duration.seconds(30),
    healthy_http_codes="200"
)

Set runtimePlatform for ApplicationLoadBalancedFargateService

# cluster: ecs.Cluster

application_load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    runtime_platform=ecs.RuntimePlatform(
        cpu_architecture=ecs.CpuArchitecture.ARM64,
        operating_system_family=ecs.OperatingSystemFamily.LINUX
    )
)

Customize Container Name for ScheduledFargateTask

# cluster: ecs.Cluster

scheduled_fargate_task = ecs_patterns.ScheduledFargateTask(self, "ScheduledFargateTask",
    cluster=cluster,
    scheduled_fargate_task_image_options=ecsPatterns.ScheduledFargateTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        container_name="customContainerName",
        memory_limit_mi_b=512
    ),
    schedule=appscaling.Schedule.expression("rate(1 minute)"),
    platform_version=ecs.FargatePlatformVersion.LATEST
)

Customize Container Name for ScheduledEc2Task

# cluster: ecs.Cluster

ecs_scheduled_task = ecs_patterns.ScheduledEc2Task(self, "ScheduledTask",
    cluster=cluster,
    scheduled_ec2_task_image_options=ecsPatterns.ScheduledEc2TaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        container_name="customContainerName",
        memory_limit_mi_b=256,
        environment={"name": "TRIGGER", "value": "CloudWatch Events"}
    ),
    schedule=appscaling.Schedule.expression("rate(1 minute)"),
    enabled=True,
    rule_name="sample-scheduled-task-rule"
)

Set PlatformVersion for ScheduledFargateTask

# cluster: ecs.Cluster

scheduled_fargate_task = ecs_patterns.ScheduledFargateTask(self, "ScheduledFargateTask",
    cluster=cluster,
    scheduled_fargate_task_image_options=ecsPatterns.ScheduledFargateTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        memory_limit_mi_b=512
    ),
    schedule=appscaling.Schedule.expression("rate(1 minute)"),
    platform_version=ecs.FargatePlatformVersion.VERSION1_4
)

Set SecurityGroups for ScheduledFargateTask

vpc = ec2.Vpc(self, "Vpc", max_azs=1)
cluster = ecs.Cluster(self, "EcsCluster", vpc=vpc)
security_group = ec2.SecurityGroup(self, "SG", vpc=vpc)

scheduled_fargate_task = ecs_patterns.ScheduledFargateTask(self, "ScheduledFargateTask",
    cluster=cluster,
    scheduled_fargate_task_image_options=ecsPatterns.ScheduledFargateTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        memory_limit_mi_b=512
    ),
    schedule=appscaling.Schedule.expression("rate(1 minute)"),
    security_groups=[security_group]
)

Deploy application and metrics sidecar

The following is an example of deploying an application along with a metrics sidecar container that utilizes dockerLabels for discovery:

# cluster: ecs.Cluster
# vpc: ec2.Vpc

service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    vpc=vpc,
    desired_count=1,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        docker_labels={
            "application.label.one": "first_label",
            "application.label.two": "second_label"
        }
    )
)

service.task_definition.add_container("Sidecar",
    image=ecs.ContainerImage.from_registry("example/metrics-sidecar")
)

Select specific load balancer name ApplicationLoadBalancedFargateService

# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    desired_count=1,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    task_subnets=ec2.SubnetSelection(
        subnets=[ec2.Subnet.from_subnet_id(self, "subnet", "VpcISOLATEDSubnet1Subnet80F07FA0")]
    ),
    load_balancer_name="application-lb-name"
)

ECS Exec

You can use ECS Exec to run commands in or get a shell to a container running on an Amazon EC2 instance or on AWS Fargate. Enable ECS Exec, by setting enableExecuteCommand to true.

ECS Exec is supported by all Services i.e. ApplicationLoadBalanced(Fargate|Ec2)Service, ApplicationMultipleTargetGroups(Fargate|Ec2)Service, NetworkLoadBalanced(Fargate|Ec2)Service, NetworkMultipleTargetGroups(Fargate|Ec2)Service, QueueProcessing(Fargate|Ec2)Service. It is not supported for ScheduledTasks.

Read more about ECS Exec in the ECS Developer Guide.

Example:

# cluster: ecs.Cluster

load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    desired_count=1,
    cpu=512,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    enable_execute_command=True
)

Please note, ECS Exec leverages AWS Systems Manager (SSM). So as a prerequisite for the exec command to work, you need to have the SSM plugin for the AWS CLI installed locally. For more information, see Install Session Manager plugin for AWS CLI.

Propagate Tags from task definition for ScheduledFargateTask

For tasks that are defined by a Task Definition, tags applied to the definition will not be applied to the running task by default. To get this behavior, set propagateTags to ecs.PropagatedTagSource.TASK_DEFINITION as shown below:

from aws_cdk import Tags


vpc = ec2.Vpc(self, "Vpc", max_azs=1)
cluster = ecs.Cluster(self, "EcsCluster", vpc=vpc)
task_definition = ecs.FargateTaskDefinition(self, "TaskDef",
    memory_limit_mi_b=512,
    cpu=256
)
task_definition.add_container("WebContainer",
    image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
)
Tags.of(task_definition).add("my-tag", "my-tag-value")
scheduled_fargate_task = ecs_patterns.ScheduledFargateTask(self, "ScheduledFargateTask",
    cluster=cluster,
    task_definition=task_definition,
    schedule=appscaling.Schedule.expression("rate(1 minute)"),
    propagate_tags=ecs.PropagatedTagSource.TASK_DEFINITION
)

Pass a list of tags for ScheduledFargateTask

You can pass a list of tags to be applied to a Fargate task directly. These tags are in addition to any tags that could be applied to the task definition and propagated using the propagateTags attribute.

vpc = ec2.Vpc(self, "Vpc", max_azs=1)
cluster = ecs.Cluster(self, "EcsCluster", vpc=vpc)
scheduled_fargate_task = ecs_patterns.ScheduledFargateTask(self, "ScheduledFargateTask",
    cluster=cluster,
    scheduled_fargate_task_image_options=ecsPatterns.ScheduledFargateTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        memory_limit_mi_b=512
    ),
    schedule=appscaling.Schedule.expression("rate(1 minute)"),
    tags=[Tag(
        key="my-tag",
        value="my-tag-value"
    )
    ]
)

Use custom ephemeral storage for ECS Fargate tasks

You can pass a custom ephemeral storage (21GiB - 200GiB) to ECS Fargate tasks on Fargate Platform Version 1.4.0 or later.

vpc = ec2.Vpc(self, "Vpc", max_azs=2, restrict_default_security_group=False)
cluster = ecs.Cluster(self, "FargateCluster", vpc=vpc)

application_load_balanced_fargate_service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "ALBFargateServiceWithCustomEphemeralStorage",
    cluster=cluster,
    memory_limit_mi_b=1024,
    cpu=512,
    ephemeral_storage_gi_b=21,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    )
)

network_load_balanced_fargate_service = ecs_patterns.NetworkLoadBalancedFargateService(self, "NLBFargateServiceWithCustomEphemeralStorage",
    cluster=cluster,
    memory_limit_mi_b=1024,
    cpu=512,
    ephemeral_storage_gi_b=200,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    )
)

Set securityGroups for NetworkLoadBalancedFargateService

# vpc: ec2.Vpc
# security_group: ec2.SecurityGroup

queue_processing_fargate_service = ecs_patterns.NetworkLoadBalancedFargateService(self, "Service",
    vpc=vpc,
    memory_limit_mi_b=512,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    security_groups=[security_group]
)

Set TLS for NetworkLoadBalancedFargateService / NetworkLoadBalancedEc2Service

To set up TLS listener in Network Load Balancer, you need to pass extactly one ACM certificate into the option listenerCertificate. The listener port and the target group port will also become 443 by default. You can override the listener’s port with listenerPort and the target group’s port with taskImageOptions.containerPort.

from aws_cdk.aws_certificatemanager import Certificate


certificate = Certificate.from_certificate_arn(self, "Cert", "arn:aws:acm:us-east-1:123456:certificate/abcdefg")
load_balanced_fargate_service = ecs_patterns.NetworkLoadBalancedFargateService(self, "Service",
    # The default value of listenerPort is 443 if you pass in listenerCertificate
    # It is configured to port 4443 here
    listener_port=4443,
    listener_certificate=certificate,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample"),
        # The default value of containerPort is 443 if you pass in listenerCertificate
        # It is configured to port 8443 here
        container_port=8443
    )
)

from aws_cdk.aws_certificatemanager import Certificate

# cluster: ecs.Cluster

certificate = Certificate.from_certificate_arn(self, "Cert", "arn:aws:acm:us-east-1:123456:certificate/abcdefg")
load_balanced_ecs_service = ecs_patterns.NetworkLoadBalancedEc2Service(self, "Service",
    cluster=cluster,
    memory_limit_mi_b=1024,
    # The default value of listenerPort is 443 if you pass in listenerCertificate
    # It is configured to port 4443 here
    listener_port=4443,
    listener_certificate=certificate,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("test"),
        # The default value of containerPort is 443 if you pass in listenerCertificate
        # It is configured to port 8443 here
        container_port=8443,
        environment={
            "TEST_ENVIRONMENT_VARIABLE1": "test environment variable 1 value",
            "TEST_ENVIRONMENT_VARIABLE2": "test environment variable 2 value"
        }
    ),
    desired_count=2
)

Use dualstack Load Balancer

You can use dualstack IP address type for Application Load Balancer and Network Load Balancer.

To use dualstack IP address type, you must have associated IPv6 CIDR blocks with the VPC and subnets and set the ipAddressType to IpAddressType.DUAL_STACK when creating the load balancer.

Application Load Balancer

You can use dualstack Application Load Balancer for Fargate and EC2 services.

import aws_cdk.aws_elasticloadbalancingv2 as elbv2


# The VPC and subnet must have associated IPv6 CIDR blocks.
vpc = ec2.Vpc(self, "Vpc",
    ip_protocol=ec2.IpProtocol.DUAL_STACK
)
cluster = ecs.Cluster(self, "EcsCluster", vpc=vpc)

service = ecs_patterns.ApplicationLoadBalancedFargateService(self, "myService",
    cluster=cluster,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    ip_address_type=elbv2.IpAddressType.DUAL_STACK
)

application_load_balanced_ec2_service = ecs_patterns.ApplicationLoadBalancedEc2Service(self, "myService",
    cluster=cluster,
    task_image_options=ecsPatterns.ApplicationLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    ip_address_type=elbv2.IpAddressType.DUAL_STACK
)

Network Load Balancer

You can use dualstack Network Load Balancer for Fargate and EC2 services.

import aws_cdk.aws_elasticloadbalancingv2 as elbv2


# The VPC and subnet must have associated IPv6 CIDR blocks.
vpc = ec2.Vpc(self, "Vpc",
    ip_protocol=ec2.IpProtocol.DUAL_STACK
)
cluster = ecs.Cluster(self, "EcsCluster", vpc=vpc)

network_loadbalanced_fargate_service = ecs_patterns.NetworkLoadBalancedFargateService(self, "NlbFargateService",
    cluster=cluster,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    ip_address_type=elbv2.IpAddressType.DUAL_STACK
)

network_loadbalanced_ec2_service = ecs_patterns.NetworkLoadBalancedEc2Service(self, "NlbEc2Service",
    cluster=cluster,
    task_image_options=ecsPatterns.NetworkLoadBalancedTaskImageOptions(
        image=ecs.ContainerImage.from_registry("amazon/amazon-ecs-sample")
    ),
    ip_address_type=elbv2.IpAddressType.DUAL_STACK
)