Restore Warehouse In-Place in Fabric – Step-by-Step with Restore Point
In this Microsoft Fabric tutorial, you'll learn how to restore your warehouse in-place using system or user-defined restore points. This capability is essential for disaster recovery, testing environments, or rolling back undesired changes.
🔁 What is In-Place Restore?
- Restores a warehouse to a prior point-in-time
- Overwrites the existing warehouse (name stays the same)
- Restores all components: models, queries, tables
- Useful for undoing accidental deletes, corruption, or failed deployments
📦 Types of Restore Points
- System-Created: Every 8 hours automatically (up to 180 kept)
- User-Defined: Manually created before/after major changes
- Retention for both: 30 days
✅ Common Use Cases
- Rollback after a schema error
- Reset test environments to a clean state
- Recover deleted data or fix bad ETL runs
🧭 Step-by-Step: How to Restore
- Open Microsoft Fabric Portal
- Navigate to your Warehouse
- Click Settings → Restore Points
- Select a system or user-defined restore point
- Click Restore In-Place
⚠️ If you receive Error 5064
, try again after a few moments.
🔐 Permissions Matrix
- View Restore Points: Admin, Member, Contributor, Viewer
- Create Restore Points: Admin, Member, Contributor
- Perform Restore: Admin only
💰 Billing & Cost
- Storage: Charges apply for Parquet storage under OneLake
- Compute: Charged when restore point is created or restored
- Billing is against Fabric Capacity
⚠️ Limitations
- Cannot rename warehouse on restore (same name only)
- Restore points expire after 30 days — no extensions
- Cannot restore across different warehouses
- Restore does not include downstream linked pipelines