See what drives warehouse usage on each data source
In GrowthBook 5.2, each data source has a Usage dashboard. It totals the past 30 days of queries GrowthBook ran against that warehouse and shows which experiments, fact tables, and people account for them. You can also review recent queries and get tips for reducing what your queries scan.
What the dashboard shows
- Totals for the past 30 days, broken out by experiment, fact table, person, and query type.
- Bytes scanned plus duration or slot time for BigQuery, Snowflake, and Athena. Other data sources show duration.
- A list of recent queries run against the data source.
- Optimization tips, such as turning on Data Pipeline Mode or adding a date template variable to SQL without a startDate variable.
How it fits with View Queries
The View Queries statistics cover the queries behind one analysis. The Usage dashboard looks across the whole data source instead. Use its totals to decide which experiment or fact table to investigate first.
The date template tip checks your SQL text for a startDate-based variable such as {{startDate}}. SQL template variables give your warehouse a hint about the date range a query needs. Because the check reads the SQL text, treat the tip as a prompt to review that query rather than a measured saving.
Usage collection starts with this release. The dashboard begins empty, and earlier queries are not backfilled.
Getting started
Go to Metrics and Data → Data Sources, select a data source, and open its Usage dashboard. If a tip suggests Data Pipeline Mode, note that Data Pipeline Mode is available on Enterprise plans for supported warehouses.
