THE HOSTING & INFRASTRUCTURE JOURNALISSUE 001 / NOVEMBER 2026
Website security

Website backup strategy: the 3-2-1 backup rule and restore tests

Build a website backup strategy with encrypted copies, recovery priorities and restore tests that make website backup and restore more dependable.

THE SHORT VERSION

Separate your backup copies, protect their keys and prove that you can restore critical website functions within your recovery targets.

Start your website backup strategy with recovery targets

Inventory everything needed to rebuild the site, not just its visible pages. Include the database, uploaded files, application code and configuration. Record dependencies such as DNS settings and access to deployment tools. CISA recommends identifying business-critical information, including website databases and software settings, before choosing a backup approach. [1]

Decide how much recent work you can lose and how long essential functions can remain unavailable. These decisions inform your recovery point objective, or RPO, and recovery time objective, or RTO. NIST connects recovery objectives to business impact. [2] A publishing site and a busy checkout may need different backup schedules; understanding uptime commitments and recovery objectives helps separate availability promises from recovery needs.

Apply the 3-2-1 backup rule with real separation

The 3-2-1 backup rule means three copies of important data: the working copy plus two backups, across two different media types, with one copy offsite. [3] Map those copies explicitly. Three folders on one server are not three independent recovery options, because a failure of that server's storage can remove them together. [3]

Location alone is not enough. An offsite copy that production credentials can erase still shares an access risk. CISA recommends offline backups because ransomware can target reachable copies. [4] Keep historical recovery points rather than continually replacing your only backup: rolling copies can propagate corruption. [3] Choose retention around your likely detection delay and business obligations, rather than assuming yesterday's copy will always be usable.

Encrypt copies without losing access to the keys

Backups can contain customer records and application secrets, so treat them as sensitive production data. CISA recommends encrypted, offline copies of critical information. [4] Require encrypted transfers as part of your backup design, restrict who can retrieve archives, and avoid placing downloadable backup files inside a publicly accessible website directory.

Encryption also creates a recovery dependency: you need the right key when the original environment is unavailable. NIST recommends storing keys separately from encrypted backup data while keeping them accessible for recovery. [2] Document authorized emergency access and test it. Combine this with the access controls in a secure web hosting checklist, rather than treating encryption as a substitute for permissions.

Test website backup and restore end to end

A completed backup job is not the same as a working recovery. CISA recommends scheduled recovery tests and both full and partial restores. [1] Restore into an isolated environment, disable real payments and outbound notifications, and verify pages, uploads, administrative access and database-dependent functions. Check that restored records and files belong to a consistent recovery point.

Record the backup's age, missing dependencies and elapsed time until essential functions work. Include downloading archives and obtaining keys in that measurement, not just importing files. Run exercises on a documented schedule and after material changes to the application or backup process. Assign someone to fix failures; a test that exposes a problem improves nothing until the recovery procedure changes.

Restore critical functions before secondary features

Write a recovery sequence based on business importance and technical dependencies. For example, restore trusted infrastructure and database services before the application that needs them, then verify the customer functions that matter most. CISA recommends prioritizing critical services during recovery and avoiding reinfection of clean systems. [4] After a security incident, do not reconnect a restored site before addressing the compromise.

Keep an accessible runbook that names decision-makers, identifies backup locations and explains how to validate recovery. CISA recommends documenting procedures and training the team to use them. [1] Define who approves reopening the site and who communicates any data gap. The practical goal is not restoring every feature simultaneously, but safely returning essential service within the targets your organization has agreed.

GO TO THE SOURCE

Sources & further reading

These resources provide background for the explanations above. External links are included for context, not as endorsements.

Written with AI assistance and based on the sources listed above. Published November 16, 2026. No hands-on provider testing is claimed. Read our editorial policy.

Keep this guide handy.