
In this guide: Website backups made simple: the 3-2-1 backup rule explained, what to back up, how to automate backups, and how to restore your site without panic.
Websites break. Plugins conflict, updates fail, servers crash, hackers strike, and sometimes we simply delete the wrong thing. When that happens, a usable backup can support recovery, although restoration time depends on the site and incident. No backup can mean losing years of work.
The most reliable way to think about backups is the 3-2-1 rule. Here’s what it means and how to apply it to your website.
Common situations where a backup saves the day:
Your host may keep backups, but relying on a single copy stored in the same place as your site is risky. If your hosting account is compromised or closed, those backups may disappear too.

For a typical website, that could look like:
Many professionals extend this to 3-2-1-1-0: one copy offline or immutable (can’t be modified or deleted by attackers) and zero errors after testing a restore.
A complete website backup includes two parts:
Files
wp-admin, wp-includes).wp-config.php and .htaccess.Database
The database changes more often than files, especially for stores and membership sites. An online store may need backups several times a day or real-time backups; a blog that publishes weekly can back up daily.
Also back up things outside your site: DNS records (take a screenshot or export), email (if hosted with your web host), and important account credentials stored in your password manager.
Manual backups get forgotten. Automate them:
Compare tools in Best Backup Plugins for WordPress and learn host-level setup in how to set up automatic backups on your hosting.
A backup you’ve never tested is a hope, not a plan. Here’s how to stay calm when something goes wrong:
Schedule a test restore at an interval suited to the site and after important backup-system changes. Record how long it actually takes.
Developer workflows need SSH, Git, WP-CLI, and room to run tools. A Kawedma VPS server or cloud hosting gives full control, as do an unmanaged VPS or managed VPS from Hosting.com. For client sites with staging built in, managed WordPress hosting powered by Rocket.net and Kawedma WordPress hosting are good choices, and agencies can use Kawedma reseller hosting.
Are my host’s backups enough?
They’re a great first layer, but they live with your host. An independent off-site copy protects you if the account itself has a problem.
How long should I keep backups?
Choose retention around detection time, recovery needs, storage and applicable obligations. A short rolling window may not include a usable point before a long-undetected problem.
Further reading: OWASP Top 10 web security risks
Start with the business impact of losing data. A brochure site updated once a month has different needs from a store accepting orders every few minutes. Decide how far back the business can reasonably go and how long it can operate without the website. These decisions should guide the backup frequency and recovery process.
Do not choose a schedule solely from a plugin’s default setting. List the data created by forms, orders, subscriptions and membership activity. Identify anything stored outside the main database, such as uploaded documents or a separate application. Website backups need to cover the actual system, not just the files visible in a control panel.
Write down the scope in plain language. For example, the plan may recover the WordPress files and database but not the provider’s mailboxes or a separate CRM. Knowing that boundary prevents false confidence. Add separate recovery arrangements for services outside the main archive where the business needs them.

Three folders in the same hosting account are not three independent protection layers. If the account is deleted or an attacker can erase all of them, they share a major failure mode. Compare where copies are stored and which credentials can modify or delete them. Independence matters as much as the number of archives.
A cloud destination is not automatically independent if the same compromised administrator can remove both the live site and its history. Review the available permissions and retention controls. Offline or immutable copies can add protection when configured appropriately, but understand what the product actually prevents and how recovery access is maintained.
The 3-2-1 rule is a planning principle for website backups, not proof that every risk is solved. Test whether you can still obtain a usable copy when the primary hosting dashboard is unavailable. A recovery route that requires the failed account to work has not addressed that scenario.
WordPress content and settings depend on the database, while media, themes and plugins live in files. A partial archive can leave the restored site missing important assets or configuration. Use the WordPress backup handbook to review these components and compare them with the selected tool’s documented scope.
For a changing site, the timing of database and file capture also matters. A newly uploaded product file and its database record should be recoverable together. Ask how the tool handles an active application and what consistency guarantees it provides. A large archive size alone does not demonstrate that the captured system can be restored correctly.
Keep a note of runtime and service requirements alongside website backups. A restored application may need a compatible PHP version, database engine, extensions or scheduled tasks. Configuration documentation helps recreate those dependencies without placing live secrets in an ordinary shared report.
Frequency controls how recent a recovery point can be. Retention controls how far back you can look. These solve different problems. Frequent website backups with only a very short history may be insufficient when a problem is discovered late. A long history with infrequent captures may lose too much recent activity.
Estimate storage growth and check whether the tool uses full, incremental or another supported backup method. Understand what is needed to restore a particular point. Do not delete parts of a backup chain without following the tool’s retention process, since a dependent restore point may become unusable.
Review the schedule after the site changes. Adding a store or membership feature can make the previous database frequency inadequate. More frequent capture is useful only if it completes reliably and the recovery process is understood. Record the decision and review it with the business owner.
Check the latest successful completion time, reported errors and destination usage. A task can be scheduled while failing every night. Alerts should reach someone who will act on them. Confirm that storage quotas, expired authorisation and billing issues cannot quietly stop the process for weeks.
Inspect representative archives or the tool’s validation results using the supported process. Compare expected components and investigate unusual changes in size or duration. Do not download sensitive production archives to an unmanaged shared device merely to glance at them. Backup files may contain the entire customer database and application secrets.
Keep a short operational record for website backups: last success, retention coverage, restore test date and unresolved issues. This is more useful than saving hundreds of identical success emails. The record should let another administrator understand whether the recovery plan is currently usable.
Create an isolated test environment suitable for the application. Prevent a restored store from sending real customer messages or processing live payments. Use the backup tool’s supported restoration instructions and document any manual steps. The test is meant to verify recovery without creating a second active production system.
After restoring, check representative pages, media, login and an appropriate business workflow. Inspect configuration and scheduled tasks. Record the time taken and any missing dependency. Website backups are easier to trust when the team has observed a restore working, rather than only seeing an archive listed in storage.
Secure or remove the test environment through the normal process after the exercise. A copy of production data left on an exposed staging site creates a new problem. Where practical, use redacted or synthetic data for ongoing testing while retaining protected original archives for authorised recovery.
Define who decides that a restore is needed and who performs it. Identify the recovery point, affected data and expected interruption. For a transactional site, compare the backup time with recent orders or submissions. Do not overwrite newer business activity without understanding what would be lost and how it will be reconciled.
Sometimes restoring one file or a limited component is more appropriate than replacing the whole site. Use only recovery methods supported by the tool and application, and understand dependencies before making a partial restore. A database table can relate to other data, so selective recovery may need expert review.
After production recovery, test the original symptom and the important customer journeys. Confirm the site is no longer in maintenance mode and that mail, scheduled tasks and caches behave correctly. Record what was restored and what remains unresolved. The presence of website backups does not remove the need for post-recovery verification.
If a site may be compromised, restoring yesterday’s archive can also restore the vulnerability or malicious content. Preserve relevant evidence and follow an incident response process. Identify a suitable recovery point and address the cause before relying on the site again. A backup is one recovery component, not a complete incident response plan.
Review credentials and access according to the incident findings. Keep backup storage protected so an attacker cannot simply remove every recovery option. Avoid repeatedly restoring and reconnecting an unchanged vulnerable system. If the scope is unclear, involve an appropriately qualified administrator or security specialist.
Document the lesson for future website backups. A missed alert, insufficient retention or a backup stored under the compromised account may require a process change. The useful outcome is a better recovery capability, not merely returning the homepage to its previous appearance.

If you are reviewing hosting or backup options, KAWEDMA offers hosting and CodeGuard-related backup services worth investigating. Confirm the selected product’s coverage, capture frequency, retention, storage location and restoration process. Do not assume every hosting plan includes the same backup service or an independent copy.
Ask about download access, restore fees and support response. Check how website backups remain accessible if the original hosting account is unavailable. Apply these questions to any supplier. A useful service description explains what you can recover and how, rather than only advertising that backups exist.
Not necessarily. A snapshot captures a particular infrastructure state, while an application backup may provide different export and restore options. Understand the consistency, retention and portability of the selected method. Both can be useful, but neither should be assumed to cover every recovery scenario without testing.
Avoid exposing backup archives through the public site. They can contain database records, configuration and secrets. Use storage designed for protected backups and verify access controls. Website backups should be available to authorised recovery staff, not downloadable by ordinary visitors or search engines.
Have a suitable current recovery point before changes that could affect the site. The appropriate method depends on the application and tool. WordPress recommends considering backups before updates. A small version number does not guarantee a change cannot affect a plugin interaction or customer workflow.
You know what is covered, where protected copies are held, who can access them and how a tested restoration works. Open issues are documented and the backup process is monitored. This is stronger evidence than a promise of instant recovery or a dashboard containing old archives nobody has checked.
Assign a named owner in the service agreement and give the client a concise report. Explain which data the service covers and which systems need separate protection. Website backups become a shared operational process when both parties understand the scope, rather than an assumption that the other person has a copy.
Include the tool, storage destination, schedule, retention policy, last successful restore test and authorised recovery contacts. Record where secure access information is held without copying secrets into the handover note. The next administrator should be able to follow the recovery process even when the original maintainer is unavailable.
Review website backups after changing hosts, storage destinations or major application components. Confirm that the new configuration actually runs and that old retention is handled intentionally. Do not cancel the previous backup service before the replacement has a verified usable recovery point. A brief overlap and a documented acceptance check can prevent a gap that is discovered only during an outage.
We may earn a commission on purchases made through these links, at no extra cost to you.
Don’t miss the chance to participate in the biggest giveaway of this year.