Using terraform there is a depends_on
capability for any resource type, including an ECS task-group.
ECS tasks can have a z-pages style "Health Check" behavior, when the task starts it must successfully return (usually) an HTTP 200 OK and it will be marked healthy.
Is there a mechanism where the depends_on behavior will wait/delay until there is at least one healthy task. I think right now the depends_on behavior waits until the ECS task is created, but not provisioned or running.
This would be useful in situations such as:
You can set the wait_for_steady_state = true
attribute on the ECS service Terraform definition, to get Terraform to wait for the service to be stable.