A BigQuery data mart serves many users. One user needs to run non-urgent jobs that must not impact critical workloads and is indifferent to runtime. To minimize cost and admin effort, what should you ask them to do?

Running jobs as batch jobs in BigQuery is the most cost-effective and administratively simple solution for non-urgent tasks. Batch jobs are queued and run when resources are available, typically at a lower priority than interactive queries. This ensures they don't contend with critical workloads, and because they are indifferent to runtime, the user won't be negatively impacted by the delayed execution. This approach leverages BigQuery's built-in resource management. Creating a separate project adds unnecessary administrative overhead and doesn't inherently prioritize workloads. Adding the user as a job.user role only grants permission to run jobs, not to manage their priority or resource consumption. Allowing the user to run jobs only when important workloads are not running requires manual coordination and scheduling, which is prone to error and administrative burden.