In this guide: How to set up automatic backups on your hosting: host backup options, plugin-based backups, off-site storage, and a monthly restore test.
Backups are the one thing every website owner agrees are important — and the thing many forget to set up properly. The solution is automation: assign a schedule, verify its results, and keep a tested recovery procedure.
Manual backups fail for predictable reasons:
Automatic backups run on a schedule, keep multiple versions, and can provide recovery points when successfully completed and retained — whether it’s a failed update, malware, a hosting issue, or human error. See Website Backups: The 3-2-1 Rule Explained.
Start by understanding what your hosting already provides. Check your control panel and plan details:
Common control panel tools:
Tip: create a manual/on-demand backup before every major update, even with automatic backups enabled.
A backup plugin can send a separate copy outside the hosting account when remote storage is configured and verified. Installing the plugin alone does not establish that separation.
Setup example with a typical backup plugin:
wp-content folders, and the database.For large sites, choose plugins with incremental backups or services that process backups on their own servers to reduce hosting load.
Store at least one copy completely separate from your host:
Security tips:
A backup you’ve never restored is only a hope. Test regularly:
Monthly or quarterly:
Also verify:
Further reading: WordPress Advanced Administration Handbook
Start with a business question: what information could you afford to recreate, and how long could the site remain unavailable? A brochure site with occasional edits has different needs from a store that receives orders throughout the day. Do not copy another site’s schedule without considering that difference. Write down the maximum acceptable gap in recovered data and the desired recovery window, then ask whether the proposed backup process can support them.
Automatic backups should cover the components required to rebuild the service, not merely the easiest directory to download. WordPress needs its database and relevant files. Hosting email, DNS settings, scheduled jobs, application configuration and external services may sit outside a WordPress plugin’s scope. Keep a component inventory showing which system protects each item and who can recover it. Do not store live passwords in that inventory.
A WordPress content export is useful for transferring supported content, but it should not be mistaken for a complete hosting account backup. Likewise, a provider snapshot may cover a virtual machine without guaranteeing that every application has written a consistent transaction state. Ask the provider how the snapshot is taken, what it excludes and what restore access remains if the primary account is unavailable.
Use the official WordPress backup guidance to understand the files-and-database distinction. A matching recovery set helps avoid restoring old database references against newer uploads. The exact consistency method depends on the application and backup tool. For a busy transactional site, have an experienced administrator assess the appropriate method rather than assuming that two independently scheduled copies always match.
Seeing a Backup Wizard button does not prove that automatic backups are running. A manual archive generator and a host-managed schedule are different features. Check the schedule, most recent successful completion, destination and retention in the interface available to your account. Ask support for clarification if the panel shows no schedule. Keep the response with your recovery documentation instead of guessing from a plan name.
In cPanel, a full account backup and a partial restore have different workflows. The cPanel Backup documentation explains supported operations; a normal account user should not assume that a full archive can be restored directly through every Backup Wizard screen. Confirm whether provider assistance or administrator privileges are required before relying on that archive in an emergency.
Choose database frequency based on the rate of important writes. Orders, customer registrations and form submissions may matter more than the frequency of publishing blog posts. If a daily database copy would lose too much work, investigate a more suitable service and confirm how its recovery points work. More frequent jobs are not automatically better if they repeatedly fail, overlap or overload a constrained server.
File frequency should reflect uploads, software changes and custom development. Schedule resource-intensive jobs away from known busy periods where practical, but account for the audience’s different time zones. Record the scheduler’s actual timezone and any daylight-saving effects. If the tool depends on WordPress traffic to trigger a job, review its documented scheduling mechanism with the host before treating the displayed start time as a precise guarantee.
A second archive in the same public web directory is vulnerable to many of the same failures as the live site. It may also expose sensitive content if downloadable without authorization. Use a destination outside the publicly served directory and, where appropriate, outside the primary account. Separation should consider credentials and deletion permissions as well as the physical location of storage.
Restrict the credentials used by the backup job to the supported operations it needs. A simple write-only policy may not work with a tool that must list or remove old objects, so test the documented permissions rather than applying a slogan. Protect the storage administrator account separately. Verify that an attacker controlling the website cannot trivially delete every recovery point through a broadly privileged storage key.
Retention is a business and data-governance choice, not a universal seven-day rule. Consider how long a mistake or compromise might remain unnoticed, how quickly the site changes, storage costs and any obligations applicable to the data. Keeping more versions can help, but retaining sensitive customer records indefinitely creates its own exposure. Write a policy and review it when the site’s business model changes.
Versioning and immutable storage have different capabilities and configuration requirements. Do not claim that enabling one checkbox makes all automatic backups undeletable. Confirm supported retention periods, administrative exceptions, account security and billing implications with the storage provider. Test the behavior using non-sensitive sample data before depending on it. Keep encryption keys and recovery instructions available to authorized staff through a separate protected process.
A successful job notification is one piece of evidence. Check the timestamp of the actual recovery point, the files included and any warnings about skipped paths or failed database exports. An archive with a plausible size may still be incomplete. Track the age of the newest usable recovery point against your requirement, not simply whether the plugin has a green icon.
Send failure alerts to an address monitored by someone responsible for acting on them. If that mailbox is hosted on the same unavailable server, have another authorized notification route. Do not put database passwords, customer records or complete debug dumps into routine email alerts. Record who investigates a failed job, how quickly they respond and when the issue should be escalated to the host or developer.
Run a restoration exercise in an isolated environment whose purpose is clear to the team. Prevent test copies from sending customer messages, charging payments or triggering production integrations. Use the service’s documented sandbox modes where available and verify that outgoing jobs are controlled before opening the copy. Merely adding a noindex instruction does not make a staging site private or stop its automation.
Check representative user journeys after restoration: a recent article and image, a permitted login, a form submission to a test destination, and a store checkout using an appropriate sandbox. Confirm that private data remains protected and that URLs, scheduled tasks and external integrations point to the intended environment. Document exceptions. A homepage loading successfully does not establish that every recovery component works.
Restoring an older database can overwrite newer orders, messages or account changes. Before a production rollback, preserve the current state where safe and identify what would be lost. Decide whether a selective file recovery, a forward fix or a full restore is appropriate. A store may need professional help reconciling transactions; do not merge databases casually based on a generic tutorial.
Coordinate the change with an owner who can approve the business impact. Keep a record of the chosen recovery point, expected data gap, steps performed and validation results. Prepare a fallback if the restore fails. Follow the incident communication workflow when customers are affected, stating verified facts rather than promising that every order or payment is safe before checking.
Repeat the review after a hosting migration, new payment integration, storage change or major increase in uploads. A working schedule can silently become unsuitable when credentials expire or a destination fills. Include recovery ownership in your maintenance plan and connect it with the update workflow so a pre-change copy is verified before a risky release.
For an illustrative small business, the evidence record might list the last completed database job, matching file set, destination, retention rule, alert recipient and last isolated restore exercise. These are fields to complete with your own evidence, not results measured by Hostreta. Mark unanswered items as unverified and assign an owner. This turns automatic backups into a maintained recovery process instead of an installation task that nobody revisits.
Promotional recommendation: if you need a backup service to evaluate, review KAWEDMA’s CodeGuard offering. Confirm the exact supported site, database coverage, storage, retention and restoration assistance before purchasing. This is a service recommendation, not a claim that Hostreta tested its recovery time or that every plan meets your requirements.
Affiliate disclosure: Hostreta may earn a commission if you purchase through this link. When comparing hosting that includes a managed recovery workflow, you can also review hosting.com’s managed WordPress hosting. Verify the current plan’s backup and restore terms separately; managed hosting is not a substitute for understanding your own recovery responsibilities.
Imagine a small store whose checkout stopped working after a plugin update. The owner sees a backup from the previous night and another completed immediately before the update. Neither should be selected solely because it appears first in the list. Check which files and database state each point contains, whether orders arrived afterward and whether the failure can be corrected without rolling back customer data.
Restore the preferred candidate in the controlled test environment first, if the urgency and available tooling permit it. Confirm that the checkout problem is absent and that the expected software versions are present. Compare a few relevant order identifiers against the production record without copying personal data into an unsecured worksheet. The purpose is to understand the gap, not to demonstrate that a green backup icon guarantees a lossless rollback.
If the incident began before the latest automatic backups, a recent copy may preserve the same faulty configuration or malicious file. Retention history helps only when the team can identify a suitable point and address the original cause. Record what is known and what remains uncertain. Do not tell customers that no information was affected merely because the site becomes reachable again after restoration.
Ask who can authorize a restore, how access works if the main control panel is unavailable, and whether cancellation or suspension removes stored recovery points. Confirm any assistance charges and export restrictions using the current contract rather than assumptions. Ask whether backup success alerts include skipped components and how long failed jobs remain visible. Keep this information with the recovery runbook so another authorized person can act when the usual administrator is away.
The practical acceptance condition is simple: the team understands what its automatic backups protect, can obtain the required recovery material and has evidence from an appropriate restoration exercise. Review those three conditions after every major operational change.
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.