When developing a new Cloud Dataflow pipeline, you want to restrict the processing resources during the testing phase. Which parameter should you specify when running your new Cloud Dataflow job to achieve this?