This article discusses AI Commerce 's backup process to AWS S3, how customers can request access to backups, and what responsibilities customers have if they make changes to the database. This information will help you understand how your company's data is secure and what to do in case of data loss or corruption.

How backups are performed

• AI Commerce 's database is automatically backed up every four hours to AWS S3, table by table.
• A full backup is taken once a night using RDS automation (managed by AWS). This is used, for example, to restart the entire database in another data center in the event of a major disaster.
• Backups are stored in Gzip format (compact format, which keeps storage space moderate).
• Backups are not located in the same environment as the actual database, which increases data security and reliability.
• If data corruption or loss occurs in an individual table, only the backup of that table can be restored, ensuring the preservation of other data.

Customers' right to backups

• Customers have the right to request backups of their own data, for example when they want to move their data off the platform or use the data in their own environment.
• Access to backups is not automatic by default, but is granted upon separate request at no additional charge.
• After the request, the customer will be provided with AWS CLI configuration information that can be used to download backups using the tool of their choice (e.g. in the CLI terminal or with the CyberDuck application).
• However, RDS-level automatic backups (nightly backup of the entire database) are not made directly available to the customer for security reasons.

Information security and division of responsibilities

• If the customer themselves leak or lose the keys that allow access to backups, AI Commerce does not take responsibility for compliance with GDPR regulations or other consequences.
• Some customers may have access to their own database for development purposes. In this case:
  • The customer is responsible for any database changes they make.
  • The customer provides a fixed IP address to which access is granted. Dynamic IP addresses are not supported.
  • If the actions cause application errors or other problems, AI Commerce is not responsible for correcting errors or restoring data integrity.
  • However, you can always request a backup copy from AI Commerce or perform it yourself if your credentials and rights allow it.

Step-by-step instructions for obtaining backups

1. Create a request to AI Commerce stating that you need access to your backups.
2. AI Commerce will provide you with AWS CLI configuration information and, if necessary, instructions for installing tools (e.g. CyberDuck or other SFTP/CLI application).
3. Log in to AWS S3 with the tool of your choice using the credentials and settings you received.
4. Download the backups you want locally as Gzip-compressed files.
5. Extract the files and deploy them in your desired environment or keep them safe for possible restores.

Additional tips

• We recommend that you regularly take your own backups if you are processing your data elsewhere.
• Carefully consider the need for direct database access. Making database changes yourself may lead to errors for which AI Commerce is not responsible.
• Make sure that the credentials and keys that allow you to access your backups are stored securely to avoid security risks.

Summary

AI Commerce takes care of regular backups to AWS S3 and nightly backups to RDS, minimizing the risk of data loss. Customers can request access to their own S3 backups, for example when leaving the platform or to keep data for themselves. Database-level access is possible, but it shifts responsibility for the consequences of database changes to the customer.