Resource usage is collected per worker
Run-level metrics are not available for distributed runs. Each worker’s CPU and memory are on the pipeline’s Resource Usage.