Elastic Container Service (ECS)

Get started with Elastic Container Service (ECS) on LocalStack


Amazon Elastic Container Service (Amazon ECS) is a fully managed container orchestration service provided by Amazon Web Services (AWS). It allows you to run, stop, and manage Docker containers on a cluster. ECS eliminates the need for you to install, operate, and scale your own cluster management infrastructure.

Getting Started

This guide is designed for users new to ECS and assumes basic knowledge of the AWS CLI and our awslocal wrapper script.

Start your LocalStack container using your preferred method. We will demonstrate how to create an ECS service using the AWS CLI

Create a cluster

ECS tasks and services run on a cluster. Execute the following command to create an ECS cluster named mycluster:

$ awslocal ecs create-cluster --cluster-name mycluster
{ "cluster": { "clusterArn": "arn:aws:ecs:us-east-1:000000000000:cluster/mycluster", "clusterName": "mycluster", "status": "ACTIVE", "registeredContainerInstancesCount": 0, "runningTasksCount": 0, "pendingTasksCount": 0, "activeServicesCount": 0, "settings": [ { "name": "containerInsights", "value": "disabled" } ] } }

Create a task definition

Containers within tasks are defined by a task definition that is managed outside of the context of a cluster. To create a task definition that runs an ubuntu container forever (by running an infinite loop printing “Running” on startup), create the following file as task_definition.json:

  "containerDefinitions": [
      "name": "server",
      "image": "ubuntu",
      "cpu": 10,
      "memory": 10,
      "command": [
        "while true; do echo running; sleep 1; done"
      "essential": true,
      "logConfiguration": {
        "logDriver": "awslogs",
        "options": {
          "awslogs-create-group": "true",
          "awslogs-group": "myloggroup",
          "awslogs-stream-prefix": "myprefix",
          "awslogs-region": "us-east-1"
  "family": "myfamily"

and then run the following command:

$ awslocal ecs register-task-definition --cli-input-json file://task_definition.json
{ "taskDefinition": { "taskDefinitionArn": "arn:aws:ecs:us-east-1:000000000000:task-definition/myfamily:1", "containerDefinitions": [ { "name": "server", "image": "ubuntu", "cpu": 10, "memory": 10, "portMappings": [], "essential": true, "command": [ "sh", "-c", "while true; do echo running; sleep 1; done" ], "environment": [], "mountPoints": [], "volumesFrom": [], "logConfiguration": { "logDriver": "awslogs", "options": { "awslogs-create-group": "true", "awslogs-group": "myloggroup", "awslogs-stream-prefix": "myprefix", "awslogs-region": "us-east-1" } } } ], "family": "myfamily", "networkMode": "bridge", "revision": 1, "volumes": [], "status": "ACTIVE", "placementConstraints": [], "compatibilities": [ "EXTERNAL", "EC2" ], "registeredAt": 1713364207.068659 } }

Task definitions are immutable, and are identified by their family field, and calling register-task-definition again with the same family value creates a new version of a task definition.

This task definition creates a CloudWatch Logs log group and log stream for the container so you can view the service logs.

Launch a service

Finally we launch an ECS service using the task definition above. This will create a number of containers in replica mode meaning they are distributed over the nodes of the cluster, or in the case of Fargate, over availability zones within the region of the cluster. To create a service, execute the following command:

$ awslocal ecs create-service --service-name myservice --cluster mycluster --task-definition myfamily --desired-count 1
{ "service": { "serviceArn": "arn:aws:ecs:us-east-1:000000000000:service/mycluster/myservice", "serviceName": "myservice", "clusterArn": "arn:aws:ecs:us-east-1:000000000000:cluster/mycluster", "loadBalancers": [], "serviceRegistries": [], "status": "ACTIVE", "desiredCount": 1, "runningCount": 1, "pendingCount": 0, "launchType": "EC2", "taskDefinition": "arn:aws:ecs:us-east-1:000000000000:task-definition/myfamily:1", "deploymentConfiguration": { "deploymentCircuitBreaker": { "enable": false, "rollback": false }, "maximumPercent": 200, "minimumHealthyPercent": 100 }, "deployments": [ { "id": "ecs-svc/49976591540684372", "status": "PRIMARY", "taskDefinition": "arn:aws:ecs:us-east-1:000000000000:task-definition/myfamily:1", "desiredCount": 1, "pendingCount": 0, "runningCount": 1, "failedTasks": 0, "createdAt": 1709242525.05109, "updatedAt": 1709242525.051093, "launchType": "EC2", "rolloutState": "IN_PROGRESS", "rolloutStateReason": "ECS deployment ecs-svc/49976591540684372 in progress." } ], "events": [], "createdAt": 1709242525.051096, "placementStrategy": [], "schedulingStrategy": "REPLICA", "createdBy": "arn:aws:iam::000000000000:user/test" } }

You should see a new docker container has been created, using the ubuntu:latest image, and running the infinite loop command:

$ docker ps
CONTAINER ID   IMAGE                       COMMAND                  CREATED         STATUS                   PORTS                                                                                              NAMES
5dfeb9376391   ubuntu                      "sh -c 'while true; …"   3 minutes ago   Up 3 minutes                                                                                                                ls-ecs-mycluster-75f0515e-0364-4ee5-9828-19026140c91a-0-a1afaa9d
9967fe5300cc   localstack/localstack-pro   "docker-entrypoint.sh"   5 minutes ago   Up 5 minutes (healthy)>443/tcp,>4510-4560/tcp, 53/tcp, 5678/tcp,>4566/tcp   localstack-main

Collect container logs

To access the generated logs from the container, run the following command:

awslocal logs filter-log-events --log-group-name myloggroup --query 'events[].message'
$ awslocal logs filter-log-events --log-group-name myloggroup | head -n 20 { "events": [ { "logStreamName": "myprefix/ls-ecs-mycluster-75f0515e-0364-4ee5-9828-19026140c91a-0-a1afaa9d/75f0515e-0364-4ee5-9828-19026140c91a", "timestamp": 1713364216375, "message": "running", "ingestionTime": 1713364216704, "eventId": "0" }, { "logStreamName": "myprefix/ls-ecs-mycluster-75f0515e-0364-4ee5-9828-19026140c91a-0-a1afaa9d/75f0515e-0364-4ee5-9828-19026140c91a", "timestamp": 1713364216440, "message": "running", "ingestionTime": 1713364216704, "eventId": "1" }, { "logStreamName": "myprefix/ls-ecs-mycluster-75f0515e-0364-4ee5-9828-19026140c91a-0-a1afaa9d/75f0515e-0364-4ee5-9828-19026140c91a", "timestamp": 1713364216505, "message": "running",

See our CloudWatch Logs user guide for more details.

LocalStack ECS behavior

You can use the configuration option MAIN_DOCKER_NETWORK to specify the network the ECS containers are started in. Otherwise, your ECS containers will be created in the same Docker network that LocalStack is in. If your ECS containers depend on LocalStack services, your ECS task network should be the same as the LocalStack container network.

If you are running LocalStack through a docker run command, do not forget to enable the communication from the container to the Docker Engine API. You can provide the access by adding the following option -v /var/run/docker.sock:/var/run/docker.sock.

For more information regarding the configuration of LocalStack, please check the LocalStack configuration section.

Remote debugging

To enable a remote debugging port for your ECS tasks, set the environment variable ECS_DOCKER_FLAGS="-p 0:<debugger port>" to expose your debugger on a random port on your host. You can then use this port to remote attach your debugger. Or if you are working with a single container, you can set ECS_DOCKER_FLAGS="-p <debugger port>:<debugger port>" to expose the debugger port to your host system.

Mounting local directories for ECS tasks

In some cases, it can be useful to mount code from the host filesystem into the ECS container. For example, to enable a quick debugging loop where you can test changes without having to build and redeploy the task’s Docker image each time - similar to the Lambda Hot Reloading feature in LocalStack.

In order to leverage code mounting, we can use the ECS bind mounts feature, which is covered in the AWS Bind mounts documentation.

Boto3 example

The Python sample code below registers a task definition, mounting a host path /host/path into the container under /container/path:

ecs_client = boto3.client("ecs", endpoint_url="http://localhost:4566")
            "name": "...",
            "image": "alpine",
            "command": ["..."],
            "mountPoints": [
                {"containerPath": "/container/path", "sourceVolume": "test-volume"}
    volumes=[{"host": {"sourcePath": "/host/path"}, "name": "test-volume"}],

CDK example

The same functionality can be achieved with the AWS CDK following this (Python) example:

task_definition = ecs.TaskDefinition(
        ecs.Volume(name="test-volume", host=ecs.Host(source_path="/host/path"))

container = task_def.add_container(...)