Product
Data
What your stored data costs: storage by project, dataset and table, unqueried tables, and the query cost of reading each one.
The Data page (/data) answers "what does our stored data cost, and is anyone using it?" It covers every plan.
Pick a range at the top: 24H, 7D, 30D, MTD or 90D. Use the project and region pickers to narrow the scope. Press / to search for a dataset or table.
Overview
| Tile | Meaning |
|---|---|
| Storage / mo | The monthly storage bill at the last scan: active, long-term, time travel and fail-safe, at list price |
| Storage in range | Storage cost that built up over the range: each day's monthly rate ÷ the days in that month, added up |
| Read query cost | List-price compute of jobs in the range that read at least one table here |
| Logical data | Logical bytes (what you query), with physical bytes and the compression ratio |
| Tables | Tables, datasets and views |
| Unqueried / mo | Storage of tables that no job read in the scan window. Candidates to archive or drop |
Where the cost sits
Storage split by project and by cost center at the last scan, and the datasets that grew most in the range.
Tables to look at
- The most expensive tables.
- Tables no job read in the scan window.
- The datasets that queries read most.
Datasets
Every dataset in scope, most expensive storage first. Click one to open its page.
Dataset page
/data/dataset?project=…&dataset=…
- Overview tiles for the dataset.
- Charts: stored bytes, query cost per day, jobs per day.
- Tables: every table and view in the dataset. Click one to open it.
- Open findings for this dataset.
Table page
/data/table?project=…&dataset=…&table=…
- What the table stores and costs.
- Size (last 7 days): daily logical size, for tables of 1 GiB or more.
- Who reads it: the tables, people and queries behind its query cost.
- Details: catalog facts, partitioning, clustering and schema from the last scan.
- Lineage: upstream and downstream counts, with a link to Lineage.
- Open findings on this table. Each opens in a new tab.
How numbers are counted
- Storage is a daily snapshot. A month's rate is shown per day as rate ÷ days in that month.
- Query cost reading data is the list-price cost of jobs that read the dataset or table. A job that reads two datasets counts in both dataset pages, but once in the overview total.
- Cache hits cost nothing.
- Under a cost-center filter, storage shows the current snapshot only. Per-table storage history is kept 7 days, so it cannot be split over time.