Run subgroup cleanup in parallel
Now that we run resource cleanup scripts once a week instead of every day, due to various factors depending on the week, there can be many more subgroups that need to be deleted which causes the job to time out. We need to run each top level sandbox group in parallel to prevent this.