Backup, redundancy and disaster recovery

Source: Storage & Data Standards, sys_kb_id=fdc09a4d93548f908037f8bd1dba10ed.

Backup

Control	Requirement
Coverage	All critical data services
Production standard	Immutable + air-gapped
Full backup	Weekly
Incremental	Daily differential or incremental
SQL full recovery model	15-minute transaction logs
Copies	Two — primary + secondary
Retention	30 days
Blob soft delete	Recommended

[TBD: the standard does not define "critical data services" — which services qualify is not enumerated.] [TBD: the standard does not state where the secondary copy must live (region, subscription or tenant).]

Commvault is the backup tool monitored under the Monitoring & Alerting standard, and appears in its list of in-scope systems.

Redundancy

Choose from LRS, ZRS, GRS, GZRS, aligned to criticality and to the application’s RPO and RTO.

Option	Scope
LRS	Locally redundant
ZRS	Zone redundant
GRS	Geo redundant
GZRS	Geo-zone redundant

[TBD: the standard does not map criticality levels or RPO/RTO thresholds to specific redundancy options. The choice is the application owner's, aligned to their documented RPO/RTO.]

Disaster recovery

RPO and RTO documented per application, by the owner.
Failover tested at least annually.

This dovetails with the Azure Environment Standards: DR is not required for Sandbox, Dev or Test; it is optional for Stage; and it is required with a set RPO and RTO for Production.

[TBD: the standard does not state target RPO/RTO values, nor what evidence an annual failover test must produce.]

Delivery and exceptions

All storage via IaC. Terraform, approved modules only.
Exceptions require all three:
1. a documented risk assessment,
2. compensating controls,
3. EA Team approval.

Source of truth: plugins/patterson-engineering/skills/storage-data-standards/references/backup-and-dr.md in the patterson-corp repository.