AWS ECS tasks was stopped with reason The Service Discovery instance could not be registered

Question

I'm trying to setup an AWS ECS service with service discovery using ansible using the following AWS ECS task definition:

- name: (define) Create gRPC AWS task definition
  ecs_taskdefinition:
    family: "my-grpc"
    region: us-west-1
    network_mode: bridge
    state: present
    execution_role_arn: "arn:aws:iam::495388981531:role/ecs-runtime-agent"
    force_create: yes
    containers:
    - name: "my-grpc"
      image: "my-custom-grpc-image"
      essential: true
      memoryReservation: 256
      portMappings:
        - containerPort: 9000
      healthCheck:
        command: 
          - "CMD-SHELL"
          - "grpc_health_probe -addr=localhost:9000"

and the following AWS ECS service configuration:

- name: (deploy-commit) Start grpc service
  ecs_service:
    name: grpc
    cluster: "my-cluster"
    state: present
    launch_type: EC2
    force_new_deployment: yes
    region: us-west-1
    task_definition: "my-grpc"
    scheduling_strategy: REPLICA
    desired_count: 1
    deployment_configuration:
      deployment_circuit_breaker:
        enable: true
        rollback: true
      maximum_percent: 200
      minimum_healthy_percent: 100
    service_registries:
      - registryArn: "my-aws-cloud-map-service-discovery-arn"
        containerName: "my-grpc"
        containerPort: 9000

I've also created the service discovery service through the AWS web console beforehand.

When I run it, the service running on the container was able to run with the health check succeeding, but the ECS task was stopped with the following stopped reason:

The Service Discovery instance could not be registered.

What's wrong here? I can't find any way to debug this.

Answer 1

It turns out I did not enable DNS queries on the AWS Cloud Map Namespace. After enabling it, service discovery is working and the ECS service ran successfully.