Skip to main content

Restoring data

A restore has four stages: pick a snapshot → build the scope → review the pre-flight → launch and monitor. Nothing writes to Salesforce until you click Start restore on the pre-flight page.

1. Pick a snapshot​

Go to Restore & Seeding and click Start a restore (or click Restore directly on a backup run's page). In the Choose a snapshot dialog, pick the backup and the run — newest first; every successful run is an independent point-in-time snapshot.

If the snapshot isn't prepared yet, bluefactory asks you to prepare the objects you want to restore first — restores run from a prepared copy so they're fast and reliable. See Preparing snapshots.

[SCREENSHOT REQUIRED: Snapshot picker dialog with backup and run selectors]

2. Build the scope​

The restore designer opens with your snapshot's objects. Add a root object — the record set you want to recover — then tick relationships to attach related objects: parents reconnect lookups, children bring dependent records along. Filter each object down to just the records you need.

The designer is the shared Data Studio scope builder — see Building a scope for everything it can do, including restore-specific tools: Anonymize, Data mapping, Ignore fields, Match by external ID, and per-object load options.

Two useful buttons as you work:

  • Preview — opens an object's matching records in the Query Console so you can eyeball exactly what will load.
  • Save draft — stores your scope so you can come back later; drafts appear in the restore list with a Draft status.

When you're ready, click Review & launch. (A restore runs one root object tree per job — the designer will tell you if your scope has more than one root.)

[SCREENSHOT REQUIRED: Restore designer with a root object and related objects on the canvas]

3. The pre-flight​

The pre-flight page is your final review:

  1. Job name — name the restore.
  2. Target org — where records land. Sandboxes are listed first, and the target never silently defaults to production.
  3. Load mode — how records are written, for the whole job:
    • Insert — always create new records.
    • Upsert — match existing records by Id, insert the rest.
    • Update — modify matched records only, never insert. (When restoring into a different org, Update can't match by Id — the page guides you to Upsert or per-object match fields instead.)
  4. Scope review — every object with its level in the tree, filters, and record counts. On a prepared snapshot the counts are exact — the records each object will actually load, narrowed by filters and parent relationships.
  5. Data mapping & field changes — a read-only recap of every transformation rule that will apply.
  6. Acknowledgement — if the target is not a sandbox, you must explicitly confirm you understand the job writes into a live org.

Click Start restore.

[SCREENSHOT REQUIRED: Restore pre-flight with target org, load mode, and scope review table]

4. Monitor the job​

The job page shows a live progress ring and, per object: records retrieved, inserted, failed, and excluded. Expand an object to see the record-by-record mapping from source Id to destination Id — destination Ids link straight into the target org. Records excluded by a forget request are labeled as such.

When the job finishes you'll see Success, Completed with warnings (some records failed — each failure shows Salesforce's error), or Failed.

[SCREENSHOT REQUIRED: Restore job page with per-object results expanded]

Re-running and cloning​

A launched restore can't be edited — click Clone scope to start a new job from the same scope. Objects configured with a match field upsert in place, so re-runs don't create duplicates.