BigQuery on-demand pricing charges based on the amount of data scanned by each query, so uncontrolled or inefficient queries from many users can quickly drive up costs. Google recommends setting custom query quotas at the user or project level to cap how many bytes queries can process per day (QueryUsagePerDay per project, QueryUsagePerUserPerDay per user), preventing runaway costs from any single user or project. These quotas apply only to on-demand pricing; the project default is 200 TiB per day. Alternatively, switch from on-demand to capacity-based pricing with BigQuery editions: you create slot reservations (a baseline plus a maximum slot count) and assign them to projects, so compute cost is predictable and capped no matter how many bytes queries scan. Flat-rate and flex-slot commitments can no longer be purchased since July 5, 2023. Simply splitting users, warehouses, or data copies across projects does not limit the underlying cost drivers (bytes scanned or compute usage) and instead increases operational complexity.
Community Discussion