r/aws_appautoscaling_target: Support updating max_capacity, min_capacity, and role_arn attributes

[bflad]

The RegisterScalableTarget API call supports updating attributes on existing targets.

Closes #240 and potentially #968

make testacc TEST=./aws TESTARGS='-run=TestAccAWSAppautoScalingTarget'
==> Checking that code complies with gofmt requirements...
TF_ACC=1 go test ./aws -v -run=TestAccAWSAppautoScalingTarget -timeout 120m
=== RUN   TestAccAWSAppautoScalingTarget_basic
--- PASS: TestAccAWSAppautoScalingTarget_basic (67.34s)
=== RUN   TestAccAWSAppautoScalingTarget_spotFleetRequest
--- PASS: TestAccAWSAppautoScalingTarget_spotFleetRequest (69.89s)
=== RUN   TestAccAWSAppautoScalingTarget_emrCluster
--- PASS: TestAccAWSAppautoScalingTarget_emrCluster (556.08s)
=== RUN   TestAccAWSAppautoScalingTarget_multipleTargets
--- PASS: TestAccAWSAppautoScalingTarget_multipleTargets (32.41s)
PASS
ok      github.com/terraform-providers/terraform-provider-aws/aws   725.772s

[Ninir]

Can you add the role_arn update in tests so that we ensure it's all ok?
Seems rather good for me otherwise :)

[bflad]

@Ninir updated! Thanks for noticing that. I changed the test to actually update the IAM role and it triggered an IAM eventual consistency issue, which is now fixed with retries:

        * aws_appautoscaling_target.bar: 1 error(s) occurred:

        * aws_appautoscaling_target.bar: ValidationException: Unable to assume IAM role: arn:aws:iam::123456789012:role/autoscaleroleupdatecluster-tyavrhv2bh
            status code: 400, request id: eef62b85-deb0-11e7-84a9-6b6b83f6f609

 make testacc TEST=./aws TESTARGS='-run=TestAccAWSAppautoScalingTarget'
==> Checking that code complies with gofmt requirements...
TF_ACC=1 go test ./aws -v -run=TestAccAWSAppautoScalingTarget -timeout 120m
=== RUN   TestAccAWSAppautoScalingTarget_basic
--- PASS: TestAccAWSAppautoScalingTarget_basic (91.72s)
=== RUN   TestAccAWSAppautoScalingTarget_spotFleetRequest
--- PASS: TestAccAWSAppautoScalingTarget_spotFleetRequest (59.53s)
=== RUN   TestAccAWSAppautoScalingTarget_emrCluster
--- PASS: TestAccAWSAppautoScalingTarget_emrCluster (2045.85s)
=== RUN   TestAccAWSAppautoScalingTarget_multipleTargets
--- PASS: TestAccAWSAppautoScalingTarget_multipleTargets (36.86s)
PASS
ok  	github.com/terraform-providers/terraform-provider-aws/aws	2234.006s

[radeksimko]

Thanks for adding the test.

One thing that I'd consider is to repurpose the existing resourceAwsAppautoscalingTargetCreate method (call it resourceAwsAppautoscalingTargetPut or something like that) instead of adding a new one since they seem to be doing the same thing.

It's not only going to be shorter/cleaner, but should address some confusions I noticed in the current Update implementation. It seems we'd be wiping existing max_capacity and/or min_capacity and/or role_arn if it's not being changed in a single update call.

Let me know what you think.

[radeksimko]

👍 More precise matching, yey

[bflad]

Given the large amount of headache with the newly forced IAM service roles, I'm going to resubmit this PR, likely without role_arn bits.

[ghost]

I'm going to lock this issue because it has been closed for 30 days ⏳. This helps our maintainers find and focus on the active issues.

If you feel this issue should be reopened, we encourage creating a new issue linking back to this one for added context. Thanks!