Wait for ECS service to stabilize

[stekern]

• If a task definition is updated, these changes will ensure that Terraform waits until 1) the ECS service has stabilized and 2) the target group is healthy. It uses the AWS CLI to wait for these conditions to be fulfilled, and each of the commands will time out after 10 minutes.
• An additional variable is added to allow for a customizable task deregistration delay (the default is 300 seconds), in order to speed up deployments.

The first of these checks is built-in to the AWS provider >= 3.13.0 as a parameter wait_for_steady_state in the aws_ecs_service resource, but a local provisioner can be used to support older provider versions. The target group check is not supported by the AWS provider.

Any thoughts on this functionality?