Branched Storage on BigQuery
Keboola is bringing branched storage to all BigQuery projects — the same capability Snowflake projects already have. This prepares BigQuery projects for Branches 2.0, an upcoming feature with approval workflows, conflict resolution, and storage schema change tracking.
What's Happening
Branched storage gives each development branch its own isolated storage environment, separate from production. This ensures branch work never interferes with your main pipeline's data.
Key points:
- August 12, 2026: Rollout begins — projects are enabled gradually
- New BigQuery projects get branched storage from creation
There is no migration deadline and no need to delete active branches — projects with existing development branches are migrated in place and keep working.
Projects that already have development branches
Existing branches are migrated in place. Nothing is deleted, no configuration
changes, and no action is required from you.
One detail worth knowing: buckets created in a branch before the migration keep their original physical storage location. Inside the branch nothing changes — they are listed there, and reading and writing work as before. In read-only storage they stay exposed the same way production buckets are, so they remain visible from production and from other branches. This is not a new behavior — it is how those buckets already work today, and the migration does not move them.
Buckets created in a branch after the migration are fully isolated — only workspaces in that same branch can read them. Both kinds can therefore exist side by side in the same branch.
Why This Matters
For your current workflow:
- Clean separation between branch and production data
- Eliminates risk of accidental data overlap during development
For what's coming (Branches 2.0):
- Approval workflows for merge requests
- Better conflict resolution tools
- Storage schema change tracking
Questions?
Reach out to your Customer Success Manager or our Support team.