BigQuery Cost Tools

3 free tools · no login · nothing uploaded

BigQuery is unusual among cloud services in that the thing you pay for — bytes scanned, or slot-time — is not the thing you write. A query that looks cheap can read a hundred times more data than it returns, and a partition filter that looks like it prunes can silently fail to.

That gap between what a query says and what it costs is where these tools live. The query analyzer reads the structure of a statement and reports what drives its cost: the columns actually read, the pruning that works and the pruning that only appears to, and the ten-megabyte-per-table floor that makes small queries cost more than their size suggests. Without a schema it prints no dollar figure at all, because a made-up byte count is worse than an honest gap.

The slot advisor answers the other BigQuery question — whether Editions capacity would beat on-demand — by simulating your real hourly demand against every baseline increment, refilling the idle hours the extraction query does not return, and charging them. Baseline slots bill through quiet hours; a model that skips them overstates savings, which is the easiest way to get this decision wrong.

The query pack is the third piece: copy-paste SQL for the cost questions people ask during an incident. Most of it is free to run, because metadata queries scan no billable bytes.

3 tools in bigquery cost

Other categories

A query at a time, or every query continuously.

Finitizer profiles your whole BigQuery estate — the queries driving scan volume, the tables charging long-term storage at active rates, the reservations that stopped fitting — and ships an AI-rewritten cheaper version of each expensive query.