Reliable website backups: what needs to be saved?

Identify everything needed for recovery

Restoring a website may require code files, the database, user uploads and configuration. A Git repository stores committed code but does not automatically include the production database, email or environment variables. Identify these before choosing a backup method.

Make a practical plan

1. List the website's data sources and where each is stored. Include external services.
2. Decide how much new data can be lost. A presentation website updated daily and a shop receiving orders continuously need different backup frequencies.
3. Check backup availability, retention and the restore process for your hosting plan. Do not assume every service includes the same backup arrangements.
4. Where needed, keep a separate copy outside the original service. Protect it with encryption and restricted access.
5. Create a restore point before updates, migrations or other major changes. Record its time and purpose.

Test restoration

Restore a copy into a separate test environment and verify files, login and database contents. Prevent the test environment from sending real order emails or processing payments. Document the recovery steps and required permissions.

If restoring production

Check what the restore overwrites and which new data has been created since the backup. Save required newer data first. Rolling back a code deployment generally does not reverse database changes. If backup contents or restoration scope are unclear, establish them with support before an overwrite restore.