AWS ECS vs Azure Container Instances: Cuándo Usar Cada Uno (Guía Completa 2026)

1{
2  "family": "mi-api",
3  "networkMode": "awsvpc",
4  "requiresCompatibilities": ["FARGATE"],
5  "cpu": "512",
6  "memory": "1024",
7  "executionRoleArn": "arn:aws:iam::123456789:role/ecsTaskExecutionRole",
8  "containerDefinitions": [
9    {
10      "name": "api",
11      "image": "123456789.dkr.ecr.us-east-1.amazonaws.com/mi-api:latest",
12      "portMappings": [
13        {
14          "containerPort": 8080,
15          "protocol": "tcp"
16        }
17      ],
18      "environment": [
19        { "name": "SPRING_PROFILES_ACTIVE", "value": "prod" },
20        { "name": "DB_HOST", "value": "mi-rds.cluster-xxx.us-east-1.rds.amazonaws.com" }
21      ],
22      "logConfiguration": {
23        "logDriver": "awslogs",
24        "options": {
25          "awslogs-group": "/ecs/mi-api",
26          "awslogs-region": "us-east-1",
27          "awslogs-stream-prefix": "ecs"
28        }
29      }
30    }
31  ]
32}

1# Registrar task definition
2aws ecs register-task-definition --cli-input-json file://task-def.json
3
4# Crear servicio con 3 réplicas y balanceo de carga
5aws ecs create-service \
6  --cluster mi-cluster \
7  --service-name mi-api-service \
8  --task-definition mi-api \
9  --desired-count 3 \
10  --launch-type FARGATE \
11  --network-configuration "awsvpcConfiguration={subnets=[subnet-abc123],securityGroups=[sg-abc123],assignPublicIp=ENABLED}" \
12  --load-balancers "targetGroupArn=arn:aws:elasticloadbalancing:us-east-1:123456789:targetgroup/mi-api-tg/abc123,containerName=api,containerPort=8080"

1# Crear un container group con una API Spring Boot
2az container create \
3  --resource-group mi-rg \
4  --name mi-api \
5  --image miregistro.azurecr.io/mi-api:latest \
6  --cpu 1 \
7  --memory 2 \
8  --ports 8080 \
9  --ip-address Public \
10  --registry-login-server miregistro.azurecr.io \
11  --registry-username miregistro \
12  --registry-password $ACR_PASSWORD \
13  --environment-variables \
14    SPRING_PROFILES_ACTIVE=prod \
15    DB_HOST=mi-db.postgres.database.azure.com \
16  --restart-policy Always
17
18# Verificar estado
19az container show --resource-group mi-rg --name mi-api --query "{Status:instanceView.state,IP:ipAddress.ip,FQDN:ipAddress.fqdn}" --output table

1apiVersion: "2021-09-01"
2name: mi-api-group
3location: eastus
4properties:
5  osType: Linux
6  restartPolicy: Always
7  ipAddress:
8    type: Public
9    ports:
10      - protocol: TCP
11        port: 8080
12  containers:
13    - name: api
14      properties:
15        image: miregistro.azurecr.io/mi-api:latest
16        resources:
17          requests:
18            cpu: 1
19            memoryInGb: 2
20        ports:
21          - port: 8080
22        environmentVariables:
23          - name: SPRING_PROFILES_ACTIVE
24            value: prod
25          - name: DB_PASSWORD
26            secureValue: "mi-password-secreto"
27    - name: sidecar-logger
28      properties:
29        image: fluent/fluentd:v1.17
30        resources:
31          requests:
32            cpu: 0.5
33            memoryInGb: 0.5

1# Desplegar desde el YAML
2az container create --resource-group mi-rg --file deploy-aci.yaml

1# Estructura típica
2Internet → ALB → ECS Service (3 réplicas) → RDS/DynamoDB
3                   ↓
4              Auto Scaling (CPU > 70%)
5              Blue/Green Deploy via CodeDeploy

1# Flujo típico
2Azure Function (timer: 0 2 * * *) → Crea ACI → Procesa datos → Termina → Se destruye
3                                                    ↓
4                                              Azure Blob Storage

1# Si necesitas el equivalente a ECS en Azure:
2Internet → Azure Container Apps (basado en ACI + KEDA + Envoy)
3              ↓
4         Auto-scaling, Revisions, Dapr integration, Ingress built-in

1resource "aws_ecs_service" "api" {
2  name            = "mi-api"
3  cluster         = aws_ecs_cluster.main.id
4  task_definition = aws_ecs_task_definition.api.arn
5  desired_count   = 3
6  launch_type     = "FARGATE"
7
8  network_configuration {
9    subnets          = var.private_subnets
10    security_groups  = [aws_security_group.ecs.id]
11    assign_public_ip = false
12  }
13
14  load_balancer {
15    target_group_arn = aws_lb_target_group.api.arn
16    container_name   = "api"
17    container_port   = 8080
18  }
19
20  deployment_controller {
21    type = "CODE_DEPLOY" # Blue/Green
22  }
23}

1resource "azurerm_container_group" "api" {
2  name                = "mi-api"
3  location            = azurerm_resource_group.main.location
4  resource_group_name = azurerm_resource_group.main.name
5  os_type             = "Linux"
6  restart_policy      = "Always"
7  ip_address_type     = "Public"
8
9  container {
10    name   = "api"
11    image  = "miregistro.azurecr.io/mi-api:latest"
12    cpu    = 1
13    memory = 2
14
15    ports {
16      port     = 8080
17      protocol = "TCP"
18    }
19
20    environment_variables = {
21      SPRING_PROFILES_ACTIVE = "prod"
22    }
23  }
24
25  image_registry_credential {
26    server   = "miregistro.azurecr.io"
27    username = var.acr_username
28    password = var.acr_password
29  }
30}