Compute model adjustments

BigQuery queries can use different compute models based on:

• On-demand (default): Billed per data processed

• Project-level assignment: Billing project assigned to a reservation

• Query-level assignment: Query with specific reservation (overrides project settings)

These levels provide flexibility to optimize compute models across your workloads.

Pipeline Level Assignment

Masthead optimizes compute at the pipeline level, delivering significant savings with minimal effort. Ready-to-use solutions are available for common orchestration technologies to assign compute models across all jobs.

dbt

Use the official dbt-reservations package for model-level reservation assignment.

Dataform

Use the official dataform-reservations package for action-level reservation assignment.

Airflow

reservation_id = 'RESERVATION_ID'
sql_with_overridden_reservation = f"""SET @@reservation='{reservation_id}';
{your_sql}
"""
BigQueryInsertJobOperator(
    task_id="run_query",
    configuration={
        "query": {
            "query": sql_with_overridden_reservation
        }
    },
    ...
)

Project Level Assignment

For tools that run BigQuery jobs without job-level compute control (e.g., Looker), use a dedicated project and a project-level assignment.

Migration steps:

1. Assign the project to a reservation with job type QUERY

2. Grant BigQuery Job User role to the service account

3. Update Billing Project ID in the tool's BigQuery connection (if unavailable, adjust Project ID instead)

4. Verify functionality

Fully qualified references for datasets and tables are required (starting with project_id where the data resides). Otherwise BigQuery will be looking for the table in the billing project.

Done! Queries from this tool now use the assigned reservation capacity.