# Disaster Recovery

Disaster recovery prepares an organization to restore critical technology services and data after a severe, prolonged disruption.

- Type: practice
- Domain: Engineering & architecture
- Level: Advanced
- Last reviewed: 2026-09-22
- Original: [Soroush’s Product Wiki](<https://wiki.soroush.uk/topic/disaster-recovery/>)

## What it means

Disaster recovery, or DR, is the set of plans, architectures, responsibilities, and tested procedures used to restore technology after events that normal high-availability mechanisms cannot absorb. It addresses priorities, dependencies, alternate environments, backups, communications, and two central targets: recovery time and acceptable data loss.

## Why it matters

Major regional outages, destructive changes, cyber incidents, and facility loss can exceed routine incident procedures. Recovery decisions made for the first time under pressure are slow and error-prone. DR planning makes business priorities explicit and tests whether backups, access, capacity, dependencies, and people can support restoration.

## When to use it

Maintain DR capability for services whose prolonged loss or data corruption would create unacceptable business, safety, regulatory, or customer consequences. The strategy should follow a business impact analysis. Not every workload needs active-active infrastructure; the recovery approach should match its required recovery time objective and recovery point objective.

## How it works

1. Identify critical services, data, dependencies, owners, and acceptable downtime and data loss.
2. Choose recovery patterns and alternate resources that can meet those objectives.
3. Protect backups and recovery credentials from the same failures as production.
4. Document declaration, communication, restoration order, validation, and return-to-normal procedures.
5. Exercise the plan regularly, record actual recovery performance, and correct gaps.

## A practical example

A payroll platform targets restoration within four hours with no more than fifteen minutes of transaction loss. It maintains encrypted cross-region backups and infrastructure definitions, rehearses recovery in an isolated account, validates balances before reopening access, and records that a manual identity dependency threatens the target.

## Where it breaks

A written plan is not demonstrated recoverability. Backups may be corrupt, dependencies unavailable, or restored data inconsistent. Lower recovery times and data-loss targets generally cost more and add complexity. Exercises can also disrupt teams, so scope and cadence should reflect risk without becoming a checkbox ritual.

## Common mistakes

- Assuming high availability or backups alone constitute disaster recovery.
- Setting recovery objectives without business impact or dependency analysis.
- Testing restoration components separately but never exercising the end-to-end service.

## Visualization

[View the visualization on the website](<https://wiki.soroush.uk/topic/disaster-recovery/#visualization-disaster-recovery>)

## Sources

1. [Contingency Planning Guide for Federal Information Systems](<https://csrc.nist.gov/pubs/sp/800/34/r1/upd1/final>) — National Institute of Standards and Technology · Government guidance
2. [Disaster Recovery of Workloads on AWS](<https://docs.aws.amazon.com/whitepapers/latest/disaster-recovery-workloads-on-aws/disaster-recovery-dr-objectives.html>) — Amazon Web Services · Cloud architecture whitepaper

## Related topics

- [Incident Management](<https://wiki.soroush.uk/topic/incident-management/>)
- [Chaos Engineering](<https://wiki.soroush.uk/topic/chaos-engineering/>)
- [Observability](<https://wiki.soroush.uk/topic/observability/>)
