Before trusting it, ask one question
If your website disappeared today, what would you need to get it running again? An existing backup is a good start. On its own, it does not tell you whether everything is included, whether someone can access it when needed or whether it can be restored. Use this practical checklist to talk to your provider.
1. Define what needs to be recovered
A website may depend on files, a database, configuration and external services. A file backup may leave out orders or content stored in a database. Inventory the necessary parts and identify who is responsible for each. Keep passwords out of shared documents: document how access can be recovered through a secure channel.
2. Agree on frequency and retention
Frequency determines how much change might be missing from the latest backup. Retention determines how far back you can go if a problem is discovered late. These are separate decisions. An information website and an application with daily transactions may need different arrangements. Set them out in the proposal and check how backup failures are handled.
3. Keep the backup outside the incident
Ask whether the same mistake, compromised account or infrastructure problem could affect both the original and its backups. Review the destination, permissions and who can delete versions. Cloud does not automatically mean independent, immutable or verified. Those properties need to be specified and checked rather than assumed to be included.
4. Restore into an isolated environment
A useful check recovers a backup without overwriting production. Check that the application starts, expected data is present and relationships remain consistent. Block email, payments and automated jobs during this test. Record the date, the backup used, the outcome and any issues; one successful test does not verify every future backup.
5. Write down the way back
A short procedure should identify who decides to restore, who carries it out and how the result is checked. Recovery time and acceptable data loss need to be agreed around the business; they are not implicit service promises. Review the procedure whenever the website, provider or dependencies change.
What you can do today
Ask when the last verified restoration happened, what it covered and what remained unresolved. If there is no evidence, start by agreeing on a safe test. The SAV team can help define what needs protection and prepare a backup proposal with an explicit scope. This guide does not describe features included in the starting hosting price.