3-2-1 Backup Rule & RTO/RPO: Site Resilience Implementation Guide

Successful backups don't guarantee recoverability. This guide breaks disaster recovery into four actionable steps: data tiering, the 3-2-1 strategy, clear RTO/R…

Insights:3-2-1 Backup Rule & RTO/RPO: Site Resilience Implementation Guide

"We back up every day" — that won't save you when things actually go wrong. The most common scenario in the industry: the backup files exist, but no one knows where they are, what point in time they can restore to, or whether restoring them will overwrite newer data.

Disaster recovery comes down to two numbers. Figure those out first, then talk about tools:

  • RPO (Recovery Point Objective): How much data loss can you tolerate? A display website might accept losing a day of data, but an inquiry form or order system cannot;
  • RTO (Recovery Time Objective): How long can you afford to be down? Four hours of downtime might mean a few lost inquiries for a website, but for an online business it's direct revenue loss.

Layer your data first — don't treat everything equally

Treating all files as the same type of data is the number one reason backup plans fail. We recommend organizing into three layers based on impact:

  • High-frequency backup: Databases (inquiries, subscriptions, articles), admin accounts and configurations;
  • Periodic backup: Application code, uploaded images and attachments, certificate files;
  • Archive storage: Historical versions, old campaign pages, legal archive materials.

After layering, you'll find that only the database may truly need hourly backups — everything else can be daily. This alone saves significant storage and validation costs.

The 3-2-1 Rule: Three Copies, Two Media Types, One Offsite

This classic rule remains relevant because it addresses "single points of failure"—whether a hard drive fails, a cloud account is compromised, or a data center floods, no single incident should cause all your data to vanish simultaneously.

  • 3 copies: production data + local backup + offsite backup;
  • 2 media types: such as local disk + object storage—don't put both on the same physical drive;
  • 1 offsite copy: at least one copy outside the same server and account.

The most common mistake is "storing backups on the same server." Once an attacker gains access, deleting backups is usually the first thing they do, wiping out both you and your data in one move.

Backup Frequency Should Follow Your RPO

Daily backups + weekly full backups + monthly archiving is the most cost-effective combination for small and medium businesses. The real focus shouldn't be on how frequent the backups are, but on getting these three things right:

  • Someone must receive alerts when a backup fails (a backup without monitoring is no backup at all);
  • Old backups should be cleaned up on schedule so storage costs don't spiral silently;
  • Backup files must be encrypted, and permissions for production and backup storage should be separated to prevent a single compromised account from losing everything.

A Recovery Drill Is a Hundred Times More Important Than a Backup Success Notification

The system showing "Backup Complete" only means the files were written successfully — it does not guarantee that your business can be restored. In real incidents, common failure points include: database version mismatch, changed attachment paths, missing configuration files, unswitched domain DNS resolution, and overwritten permission files.

So a quarterly small-scale restore drill is enough: restore the latest backup to a test environment, verify that the homepage, login, listings, form submissions, image displays, and admin permissions all work correctly, and record the recovery time. The goal of the drill is not to prove "we have no issues," but to uncover those few critical break points in advance.

If you can only do one thing right now: log into your server and confirm whether the backup file lives on the same machine as your website. If so, move it today. For related infrastructure planning, see our Infrastructure Consulting service; monitoring and alerting best practices are covered in the next article.