Database branching

Branches let every engineer work with production-like data without copying terabytes of storage. Pixetric keeps a single durable storage fabric and exposes lightweight branch pointers.

Lifecycle

1. Open the Branches page and click New branch.
2. Choose a source (usually main) and name the branch after your feature or pull request.
3. Attach a compute pool so the branch can serve queries.
4. Test, run migrations, or connect CI previews to the branch.
5. When ready, promote or delete the branch from the same drawer.

Merging changes

Apply schema changes inside the branch using your preferred migration tool. When you're satisfied, click Promote to merge into the target branch. Pixetric replays WAL changes while ensuring conflicts are surfaced before they hit production.

Refreshing data

QA branches that live for multiple days can drift from production. Use the Reset branch action to fast-forward a branch to the latest snapshot of main, optionally on a schedule (e.g., nightly at 03:00 UTC).

Branch policies

Administrators can enforce read-only policies, IP allow lists, or multi-step approvals per branch. Configure policies under Settings → Policies to protect production from accidental writes while keeping development branches flexible.

Automate previews

• Integrate your CI provider so every pull request provisions a branch automatically.
• Inject branch-specific connection strings into preview deployments.
• Delete the branch once the pull request is merged to keep compute usage lean.

Database branching eliminates contention for shared staging databases and keeps development velocity high.