remove healthCheckGracePeriodSeconds validation

[ijin]

Problem

When running ECS services without load balancers, ecspresso verify fails if healthCheckGracePeriodSeconds is defined in the service definition. However, AWS ECS automatically adds this field (with value 0) to all services, making this validation unnecessary.

Current Behavior

// Service definition without healthCheckGracePeriodSeconds
{
  "availabilityZoneRebalancing": "ENABLED",
  "desiredCount": 1,
  "launchType": "FARGATE",
  "networkConfiguration": {
    "awsvpcConfiguration": {
      "assignPublicIp": "ENABLED",
      "securityGroups": [
        "sg-xxxxxxxx"
      ],
      "subnets": [
        "subnet-aaaaaa",
        "subnet-cccccc"
      ]
    }
  },
  "platformVersion": "LATEST",
  "healthCheckGracePeriodSeconds": 1,
  "propagateTags": "NONE"
}

$ ecspresso diff
-  "healthCheckGracePeriodSeconds": 0,

$ aws ecs describe-services --cluster test --services test | grep health
            "healthCheckGracePeriodSeconds": 0,

When adding the field to match AWS's behavior:

{
  "healthCheckGracePeriodSeconds": 0,
  "desiredCount": 1,
  // ... rest of service definition
}

ecspresso verify fails:

$ ecspresso verify
2025/02/17 03:50:37 test Starting verify
  TaskDefinition
    ExecutionRole[arn:aws:iam::012345678912:role/ecsTaskExecutionRole]
    --> [OK]
    ContainerDefinition[test]
      Image[redis]
      --> [OK]
      LogConfiguration[awslogs]
      --> [OK]
    --> [OK]
  --> [OK]
  ServiceDefinition
  --> [NG] service has no load balancers, but healthCheckGracePeriodSeconds is defined

Solution

Remove the validation check for healthCheckGracePeriodSeconds because:

1. ECS adds this field by default with value 0
2. Having this field is valid for services without load balancers
3. The current validation causes unnecessary friction