A backup is only useful if you can locate it, restore it, and confirm that the restored site works. This guide provides a repeatable workflow for auditing web hosting backups, testing files, databases, email, and configuration data in a safe environment, and deciding when a rollback is appropriate.
Overview
Hosting dashboards often display backup dates and retention periods, but a successful backup job does not automatically mean you have a complete disaster-recovery plan. A backup can be too old, incomplete, difficult to download, incompatible with your current software, or dependent on access to a control panel that is unavailable during an outage.
Backup verification has two separate goals:
- Confirm that recovery points exist: Identify what was backed up, when it was captured, where it is stored, and how long it will remain available.
- Prove that recovery works: Restore a copy to an isolated staging environment, inspect the result, and document the steps needed to repeat the process.
Do not treat a production restore as your first test. A test restore should use a temporary domain, staging installation, separate server, or local environment whenever possible. Keep the production site unchanged until you have verified the recovery point and understand the rollback procedure.
Before starting, record the site’s current hosting plan, control panel, application version, database engine, PHP or runtime version, DNS provider, SSL setup, scheduled tasks, and third-party integrations. These details help you distinguish a backup problem from an environment problem.
What to track
1. Backup coverage
List every component required to bring the site back online. Website files alone are rarely sufficient. Track the following separately:
- Application files, themes, plugins, uploads, media, and user-generated content.
- Databases, including the database name, engine, character set, and credentials required by the application.
- Email mailboxes, messages, aliases, forwarding rules, and address books if email is hosted with the same provider.
- DNS records, subdomains, redirects, and domain settings. DNS is often managed separately from web files and may not be included in a hosting backup.
- Configuration files, environment variables, secrets, cron jobs, scheduled tasks, rewrite rules, and deployment settings.
- SSL certificates and renewal arrangements. A certificate may be reissued, but the recovery process should specify who controls the domain and where the certificate is configured.
- Server-specific settings, such as firewall rules, PHP modules, container definitions, reverse-proxy settings, or custom deployment scripts.
Mark each item as included, exported separately, or not covered. A gap is not necessarily unacceptable, but it should have an owner and a documented replacement procedure.
2. Restore points and retention
Record the timestamp, time zone, and scope of each available restore point. Note whether backups are automatic, manual, incremental, or full. Also record the oldest available point, the frequency of new points, and whether deleted or overwritten files can be recovered.
Do not assume that a provider’s advertised retention period applies to every backup type. Control-panel snapshots, database exports, email backups, and provider-managed disaster recovery may have different schedules or access rules. Ask support for clarification when the dashboard does not show these details.
3. Provider capabilities and costs
Use this comparison table when reviewing a host or documenting your current plan. Fill it with confirmed information from the provider’s dashboard, terms, or support response rather than relying on a sales summary.
| Item to compare | What to record | Why it matters |
|---|---|---|
| Backup frequency | How often files, databases, and email are captured | Determines the maximum likely data gap |
| Retention | Number of recovery points and retention duration | Shows whether older corruption or deletion can be reversed |
| Storage location | Same server, separate storage, downloadable copy, or external destination | Reveals whether a server failure could affect both site and backup |
| Restore method | Self-service, support-assisted, command-line, or manual export | Indicates how quickly and independently you can recover |
| Supported formats | Archive, database dump, snapshot, or provider-specific format | Affects portability to another host |
| Restore fees | Any stated charge for full or partial restores | Prevents surprises during an incident |
| Access requirements | Control-panel account, support ticket, identity checks, or backup credentials | Identifies dependencies during an outage |
4. Evidence from a test restore
A backup is not verified until you can inspect the restored result. Check that the homepage loads, internal links work, images appear, users can sign in, forms submit, background jobs run, and the application can read and write to its database. For an online store, use a safe test mode and do not place real orders or send customer emails.
Test email separately. Confirm that historical messages, mailbox folders, aliases, and forwarding rules are present if they are part of your recovery requirements. A website restore may succeed while email remains unavailable because mail is handled by another system.
Cadence and checkpoints
Use a schedule that reflects the site’s change rate and business impact. A small brochure site may need a monthly test, while a frequently updated application, publication, or store may warrant a monthly restore test and more frequent backup review.
Monthly backup review
- Confirm that recent backup jobs completed and that expected restore points are visible.
- Check available storage, failed-job notices, and changes to the hosting plan or control panel.
- Download or inspect a representative file and database export where the provider allows it.
- Verify that backup alerts still reach an active mailbox or monitoring system.
Quarterly test restore
- Create or refresh an isolated staging environment.
- Restore the latest suitable backup without changing production.
- Test the application, database connections, media, authentication, forms, scheduled tasks, and email workflow.
- Measure the time from starting the restore to passing the basic acceptance checks.
- Record failures, missing components, provider responses, and corrective actions.
Test after important changes
Run an additional test before or after a major CMS update, migration, domain change, hosting-plan change, database upgrade, infrastructure redesign, or security incident. Take a fresh recovery point immediately before a high-risk change and keep it until the change has been validated.
Restore log template
Keep each test in a shared operational record using fields such as:
Date and time: Person responsible: Site and environment: Provider and plan: Backup timestamp and time zone: Backup type and format: Restore destination: Restore start and finish time: Files verified: Database verified: Email verified: Configuration and scheduled tasks verified: Application checks passed: Problems found: Provider ticket or reference: Rollback decision: Follow-up owner and deadline:
How to interpret changes
Compare each test with the previous log instead of judging the backup system from a single successful restore. Look for changes in four areas: recovery point age, restore duration, completeness, and operational effort.
A growing gap between backup timestamps may indicate a failed schedule, storage limit, disabled job, or plan change. A restore that takes longer may reflect a larger site, slower storage, a different hosting environment, or a new approval step. Investigate the reason rather than treating duration alone as a failure.
Classify findings by impact:
- Pass: The required components restored and the acceptance checks succeeded.
- Warning: The site restored, but a noncritical item was missing, manual, or undocumented.
- Failure: A required component could not be restored, the database was unusable, credentials were unavailable, or the application could not operate.
When a restore fails, preserve the evidence. Save error messages, timestamps, screenshots, archive names, and support responses. Try an earlier recovery point to determine whether the issue is limited to one backup. If the provider uses a proprietary format or requires support intervention, document an independent export strategy so that migration remains possible.
Choose rollback based on the incident, not simply on the newest backup. A recent point may contain corrupted files, compromised credentials, or accidental deletion. An older clean point may be safer, followed by a controlled re-entry of legitimate recent data. For sites with transactions or user activity, identify what data would be lost before approving the rollback.
Production rollback should have an explicit decision owner, a maintenance notice where appropriate, a copy of the current state, and a verification plan. If DNS changes are required, record the previous records and consider DNS propagation before declaring the recovery complete.
When to revisit
Schedule a recurring backup audit monthly and a hands-on restore test at least quarterly, adjusting the interval to the site’s risk and update frequency. Revisit this workflow immediately when a provider changes its control panel, retention limits, storage location, restore process, supported formats, or restore fees. Also review it after changing hosting plans, moving between shared hosting and a VPS, adding a new database, introducing a managed service, or changing DNS and email providers.
Before each review, check the provider’s current documentation and your account dashboard. Screenshots and menu names can change, so update internal instructions when the interface changes. Confirm that the person responsible for recovery still has access and that contact details for provider support are current.
Use this practical sequence for the next test:
- Choose a recent recovery point and write down its timestamp.
- Confirm that files, databases, email, DNS, and configuration requirements are covered.
- Restore to an isolated environment.
- Run functional checks and compare the result with the live site.
- Log the duration, gaps, errors, and provider assistance required.
- Assign fixes and set the next review date.
A documented test is more valuable than an unverified promise of automatic backups. For related operational planning, review the web hosting security checklist, the website migration workflow, and the guide to hosting SLAs and restore-related service expectations. Keep the restore log with those procedures and revisit it whenever the site or hosting environment changes.