Skip to main content

ECS Services

1 - What is a Service?​

An ECS Service automatically maintains the desired number of running tasks and manages the deployment of new versions.

2 - Service types​

2.1 REPLICA service​

Maintains a specific number of tasks spread across the cluster.

{
  "serviceName": "api-service",
  "desiredCount": 3,
  "schedulingStrategy": "REPLICA"
}

2.2 DAEMON service​

Runs exactly one task on each EC2 instance of the cluster.

{
  "serviceName": "monitoring-agent",
  "schedulingStrategy": "DAEMON"
}

3 - Create a Service​

3.1 Via CLI​

aws ecs create-service \
    --cluster prod-cluster \
    --service-name api-service \
    --task-definition api:3 \
    --desired-count 3 \
    --launch-type FARGATE \
    --platform-version LATEST \
    --network-configuration "awsvpcConfiguration={
        subnets=[subnet-111,subnet-222],
        securityGroups=[sg-123],
        assignPublicIp=DISABLED
    }" \
    --load-balancers "targetGroupArn=arn:aws:elasticloadbalancing:...:targetgroup/api/123,containerName=app,containerPort=8080" \
    --health-check-grace-period-seconds 60 \
    --deployment-configuration "maximumPercent=200,minimumHealthyPercent=100" \
    --enable-execute-command

3.2 Via CloudFormation​

ECSService:
  Type: AWS::ECS::Service
  DependsOn: ALBListener
  Properties:
    ServiceName: api-service
    Cluster: !Ref ECSCluster
    TaskDefinition: !Ref TaskDefinition
    DesiredCount: 3
    LaunchType: FARGATE
    PlatformVersion: LATEST
    
    NetworkConfiguration:
      AwsvpcConfiguration:
        Subnets:
          - !Ref PrivateSubnet1
          - !Ref PrivateSubnet2
        SecurityGroups:
          - !Ref ServiceSecurityGroup
        AssignPublicIp: DISABLED
    
    LoadBalancers:
      - TargetGroupArn: !Ref TargetGroup
        ContainerName: app
        ContainerPort: 8080
    
    HealthCheckGracePeriodSeconds: 60
    
    DeploymentConfiguration:
      MaximumPercent: 200
      MinimumHealthyPercent: 100
      DeploymentCircuitBreaker:
        Enable: true
        Rollback: true
    
    DeploymentController:
      Type: ECS
    
    EnableExecuteCommand: true
    
    ServiceRegistries:
      - RegistryArn: !GetAtt ServiceDiscoveryService.Arn
    
    Tags:
      - Key: Environment
        Value: Production

4 - Deployment strategies​

4.1 Rolling Update (ECS)​

4.2 Blue/Green (CodeDeploy)​

DeploymentController:
  Type: CODE_DEPLOY

# Nécessite CodeDeploy deployment group

4.3 Deployment configuration​

ParameterDescription
MaximumPercentMax % of tasks during deployment
MinimumHealthyPercentMin % of healthy tasks during deployment
DeploymentCircuitBreakerAutomatic rollback on failure
# Déploiement prudent (pas de downtime)
DeploymentConfiguration:
  MaximumPercent: 200
  MinimumHealthyPercent: 100

# Déploiement rapide (accepte downtime temporaire)
DeploymentConfiguration:
  MaximumPercent: 200
  MinimumHealthyPercent: 50

5 - Circuit Breaker​

DeploymentConfiguration:
  DeploymentCircuitBreaker:
    Enable: true
    Rollback: true

The circuit breaker detects deployment failures and performs an automatic rollback if:

  • Tasks fail to start
  • Health checks fail
  • The deployment exceeds the timeout

6 - Service Discovery​

6.1 AWS Cloud Map​

ServiceDiscoveryNamespace:
  Type: AWS::ServiceDiscovery::PrivateDnsNamespace
  Properties:
    Name: internal.local
    Vpc: !Ref VPC

ServiceDiscoveryService:
  Type: AWS::ServiceDiscovery::Service
  Properties:
    Name: api
    NamespaceId: !Ref ServiceDiscoveryNamespace
    DnsConfig:
      DnsRecords:
        - Type: A
          TTL: 60
    HealthCheckCustomConfig:
      FailureThreshold: 1

# Dans le service ECS
ECSService:
  Properties:
    ServiceRegistries:
      - RegistryArn: !GetAtt ServiceDiscoveryService.Arn

6.2 Usage​

# Les tasks s'enregistrent automatiquement
# DNS: api.internal.local

# Depuis un autre conteneur
curl http://api.internal.local:8080/health

7 - Update a service​

7.1 New Task Definition​

# Créer nouvelle révision
aws ecs register-task-definition --cli-input-json file://task-def-v2.json

# Mettre à jour le service
aws ecs update-service \
    --cluster prod-cluster \
    --service api-service \
    --task-definition api:4 \
    --force-new-deployment

7.2 Manual scaling​

# Changer le nombre de tasks
aws ecs update-service \
    --cluster prod-cluster \
    --service api-service \
    --desired-count 5

7.3 Force new deployment​

# Forcer le redéploiement (même task definition)
aws ecs update-service \
    --cluster prod-cluster \
    --service api-service \
    --force-new-deployment

8 - ECS Exec​

Shell access to running containers.

8.1 Enable​

ECSService:
  Properties:
    EnableExecuteCommand: true

# Task Role doit avoir
{
  "Effect": "Allow",
  "Action": [
    "ssmmessages:CreateControlChannel",
    "ssmmessages:CreateDataChannel",
    "ssmmessages:OpenControlChannel",
    "ssmmessages:OpenDataChannel"
  ],
  "Resource": "*"
}

8.2 Use​

# Se connecter à un conteneur
aws ecs execute-command \
    --cluster prod-cluster \
    --task abc123 \
    --container app \
    --interactive \
    --command "/bin/bash"

9 - Managing services​

# Lister les services
aws ecs list-services --cluster prod-cluster

# Détails d'un service
aws ecs describe-services \
    --cluster prod-cluster \
    --services api-service

# Arrêter un service (set desired count to 0)
aws ecs update-service \
    --cluster prod-cluster \
    --service api-service \
    --desired-count 0

# Supprimer un service
aws ecs delete-service \
    --cluster prod-cluster \
    --service api-service \
    --force

Summary​

In this chapter, we learned:

  • The types of services (REPLICA vs DAEMON)
  • Creating services via CLI and CloudFormation
  • The deployment strategies (Rolling, Blue/Green)
  • The Circuit Breaker for automatic rollbacks
  • Service Discovery with Cloud Map
  • Updating services
  • ECS Exec for debugging

Next step​

In the next chapter, we will look at Networking and Load Balancing.

→ Next chapter: Networking and Load Balancing

← Back to the table of contents