feat(aws-ecs): include GPU & ARM based ECS optimized AMI options

[mattmcclean]

---

Pull Request Checklist

• Testing
  • Unit test added (prefer not to modify an existing test, otherwise, it's probably a breaking change)
  • CLI change?: coordinate update of integration tests with team
  • cdk-init template change?: coordinated update of integration tests with team
• Docs
  • jsdocs: All public APIs documented
  • README: README and/or documentation topic updated
  • Design: For significant features, design document added to design folder
• Title and Description
  • Change type: title prefixed with fix, feat and module name in parens, which will appear in changelog
  • Title: use lower-case and doesn't end with a period
  • Breaking?: last paragraph: "BREAKING CHANGE: <describe what changed + link for details>"
  • Issues: Indicate issues fixed via: "Fixes #xxx" or "Closes #xxx"
• Sensitive Modules (requires 2 PR approvers)
  • IAM Policy Document (in @aws-cdk/aws-iam)
  • EC2 Security Groups and ACLs (in @aws-cdk/aws-ec2)
  • Grant APIs (only if not based on official documentation with a reference)

---

By submitting this pull request, I confirm that my contribution is made under the terms of the Apache-2.0 license.

[SoManyHs]

This looks great, thanks @mattmcclean! Should we also think about modifying AddCapacityOptions to include machineImage so that these AMI constructs can be used there?

[mattmcclean]

Yes, would be good to add machineImage option to AddCapacityOptions as I see it is hard coded to be the default EcsOptimizedAmi. Would we need a base class that the others implement?

[SoManyHs]

A base class for which? The Ami classes all implement IMachineImageSource, and AddCapacityOptions extends AddAutoScalingGroupCapacityOptions. The latter inherits from autoscaling.CommonAutoScalingGroupProps which does not appear to have machineImage, though the autoscaling.AutoScalingGroupProps does. So I'm thinking we would add machineImage to AddAutoScalingGroupCapacityOptions, since that would expose the property for both the default addCapacity as well as the custom one.

[eladb]

Thanks for this! Any chance you can also add a small unit test?

[mattmcclean]

@SoManyHs cool, have added the machine image parameter to AddAutoScalingGroupCapacityOptions. Not sure how to do a test as the ami id will change frequently

[SoManyHs]

Verified that running a synth on a stack with:

cluster.addCapacity('GpuAutoScalingGroup', {
      instanceType: new ec2.InstanceType('t2.micro'),
      machineImage: new ecs.EcsOptimizedAmi({
      amiType: ecs.EcsOptimizedAmiType.Gpu
      }),

resulted in this bit of CFN:

  EcsClusterDefaultAutoScalingGroupLaunchConfigB7E376C1:
    Type: AWS::AutoScaling::LaunchConfiguration
    Properties:
      ImageId: ami-043c4e6bff652b99e
      InstanceType: t2.micro
      IamInstanceProfile:
        Ref: EcsClusterDefaultAutoScalingGroupInstanceProfile2CE606B3
      SecurityGroups:
        - Fn::GetAtt:
            - EcsClusterDefaultAutoScalingGroupInstanceSecurityGroup912E1231
            - GroupId

The ImageId: ami-043c4e6bff652b99e indeed correponds to the GPU AMI for this region (us-west-2) currently published: https://docs.aws.amazon.com/AmazonECS/latest/developerguide/ecs-optimized_AMI.html

[SoManyHs]

LGTM pending addition of the unit test suggested!

[mattmcclean]

Changes have been committed including unit tests. The CodeBuild seems to be breaking but cannot see the details. Travis CI is working fine

[rix0rrr]

Can be merged as soon as the build passes.