Backups that are proven to restore
Most hosts take backups and hope. We take backups and test them: every night, every backup is restored into a sandbox to prove it actually works. Daily, weekly or monthly schedules, one-click restores down to a single file — included from £9/month.
The industry's quiet problem
A backup you've never restored is a hope, not a backup
Almost every host takes backups. Almost none of them ever restore one until a customer is already in crisis. That is exactly when people discover the archive is corrupt, the database dump was truncated, or a config change three months ago silently broke the backup job.
We built restore testing into the platform instead. Every night, an automated job takes each backup and performs a real restore into an isolated sandbox — files unpacked, database imported, result verified. If a backup can't be restored, we know within hours of it being taken, and we fix the chain before you ever need it.
- A real restore into a sandbox — not a checksum or a size check
- Runs automatically every night, for every backup
- Failures alert our engineers the same night
- The restore path itself gets exercised daily, so it works when you click it
How backups usually fail
- •The backup job silently stops after a server change — nobody notices for months.
- •The archive exists but is corrupt or incomplete — discovered only during an emergency restore.
- •The database dump was taken mid-write and won't import cleanly.
- •The restore procedure itself was never exercised, and breaks under pressure.
How we catch all four
A nightly sandbox restore exercises the whole chain end to end: the schedule ran, the archive is intact, the database imports, and the restore machinery works. If any link breaks, the test fails and we're alerted — usually before breakfast, and always before you need that backup.
Everything the backup system does
No add-ons, no per-restore fees — this is the standard backup stack on every hosted site.
Scheduled: daily, weekly, monthly
Pick the cadence per site — daily, weekly or monthly — and take manual backups on top whenever you like, for example just before a big plugin update.
Full + incremental
Full backups capture everything; incremental backups store only what changed since the last one, so frequent backups stay fast and light on your site.
One-click full restore
Roll an entire site — files and database together — back to a previous backup from the dashboard. No support ticket, no waiting.
Single-file restore
Deleted one template file? Restore just that file from any backup without touching the rest of your site.
Single-database restore
Restore only the database — useful after a bad import or a broken migration — while leaving your files exactly as they are.
Off-site copies you control
Send backup copies to your own Amazon S3 bucket, Cloudflare R2 or any SFTP server — so a copy of your data lives somewhere only you control.
Restore what broke — nothing else
Three levels of restore, all one click
Most incidents don't need a full rollback. Restoring exactly what broke means you don't lose the orders, posts or uploads that arrived since the backup was taken.
Full site restore
A botched update left the whole site broken.
Files and database roll back together to the backup you choose. The site comes back exactly as it was at that point in time.
Single-file restore
One template or config file got overwritten.
Browse the backup, pick the file, restore it in place. Every other file — and the database — stays untouched.
Database-only restore
A bad import or plugin migration mangled your data.
Restore just the database from any backup while your files, uploads and code stay exactly as they are now.
Keep a copy somewhere only you control
Trusting your host with backups is good. Holding your own copy is better. You can configure an off-site target and G7Cloud will push backup copies to it — storage that belongs to you, under your own account, independent of us.
If you ever want to leave, audit your data, or just sleep better, your backups are already sitting in your own bucket. That's what owning your data means in practice — the same reason AI Website Builder projects let you export your code.
Your credentials, your bucket, your copy. One target per site — we don't claim "redundant off-site" storage, because honesty is a feature too.
Backups FAQ
What's included, how restores work, and what restore-testing really means.
What does "nightly restore testing" actually mean?
Every night, an automated job takes each backup and performs a real restore into an isolated sandbox environment — files unpacked, database imported. If any backup fails to restore, we are alerted and investigate before you ever need that backup in anger. It is a genuine restore, not a checksum check.
How often are backups taken?
You choose per site: daily, weekly or monthly schedules, plus manual backups whenever you want one — for example before a major update. Incremental backups keep frequent schedules lightweight.
What exactly can I restore?
Three levels: the entire site (files and database together), a single file, or just the database. All three are one-click operations from the dashboard.
Can I keep a copy of my backups outside G7Cloud?
Yes. You can configure an off-site target — your own Amazon S3 bucket, Cloudflare R2, or any SFTP server — and backup copies are pushed there. That copy is in storage you own and control.
Are backups included in the price?
Yes. Scheduled backups with one-click restore are included from the Starter plan (£9/month) on every hosted site. There is no separate backup add-on and no per-restore fee.
What happens if a nightly restore test fails?
The failed test raises an alert to our team and we investigate the backup chain for that site. The point of testing every backup every night is that a bad backup gets caught within hours of being taken — not months later when you desperately need it.
Backups pair with per-minute uptime monitoring and ScaleShield edge protection on every plan. Moving from another host? We'll migrate you for free.