1. Purpose
This plan sets out how Acrux Education restores systems and data after a disruption: what protects each part of the platform, and the steps by which each part is recovered.
It sits beside the Business Continuity Plan, which covers how Acrux keeps operating and communicating during a disruption. The two documents share one set of regions, recovery objectives, roles and test schedule, and each is stated once:
Stated in the Business Continuity Plan | Section |
|---|---|
Regions, and where each region's data and backups are held | 2 |
Recovery objectives | 3 |
Detection, assessment and response | 6 |
Communication with schools and partners | 7 |
Testing schedule | 8 |
Roles and responsibilities | 10 |
2. Scope
This plan covers all production environments of the Acrux Education platform, including Acrux Mark and Acrux Analyse, in every region.
3. How the Platform Is Built for Recovery
Each region stands alone. A region has its own database, storage and services. Losing one region does not affect schools in another.
School data stays in its jurisdiction. It is never restored or served outside the jurisdiction of the region that holds it.
Every region has a backup location. Backups are held in a second location inside the same jurisdiction. Where a country has only one location, backups are held across that location's separate data centres.
The platform is rebuilt from code. Every service, and the infrastructure it runs on, is defined in source control and can be recreated from it.
Shared services hold no school data. Sign-in routing and curriculum content are shared by every region and are recovered the same way as a region.
4. What Protects Each Part
Part | Protection | How it is recovered |
|---|---|---|
Database | Daily backups, held in the region's backup location, and point-in-time recovery | Restore from point-in-time recovery or from a daily backup |
Uploaded files | Every version kept and deleted files recoverable; held in two locations inside the region's jurisdiction where two exist | Restore the earlier version |
Application services | Run across multiple zones within the region; defined in source control | Automatic within a location; rebuilt from source control if the location is lost |
Infrastructure | Defined as code | Recreated from code |
Configuration and secrets | Held in Google Cloud Secret Manager, versioned | Restored with the infrastructure |
Routing | Cloudflare | Updated to point at the recovered services |
How often each backup is taken and how long it is kept are stated once, in Section 4.2 of the Business Continuity Plan.
5. Recovery Procedures
5.1 Database
Identify the point of failure and the affected region
If data corruption: restore from point-in-time recovery
If zone failure: restore from daily backup to an alternate zone in the same location
If the region's live location is lost: restore from the backups held in the region's backup location
Verify data integrity
Resume service
5.2 Application Services
Cloud Run automatically handles zone failures within a region
If the region's live location is lost: rebuild the region's services in its backup location, inside the same jurisdiction, from source control and infrastructure as code
Update Cloudflare routing as required
Verify functionality across affected services
5.3 Uploaded Files
Identify the affected files
Restore the earlier version, or recover the deleted file within its recovery window
Verify the restored files open correctly
Where a country has only one location, service resumes when that location returns.
6. Recovery Verification
After any recovery action, the following must be verified before declaring service restored:
Database integrity and consistency check (row counts, foreign key validation, recent transaction replay)
Application health checks pass across all affected services
End-to-end smoke test of a representative user workflow (teacher login โ assessment view โ submission)
Cloudflare routing reflects the restored topology
Audit log captures the recovery event with timestamp, scope and verification outcome