AWS says data stored only in its Bahrain region cannot be recovered after war damage

On September 16, 2026, The Register reported that AWS updated its Health Dashboard to say that resources and data hosted exclusively in its Bahrain region remain inaccessible and cannot be restored. AWS said the damage "exceeded what our regional and multi-AZ services are designed to withstand." One Bahrain Availability Zone was damaged in March and further attacks in April took down the whole region. In the UAE region, AWS said it cannot restore access to resources in one zone, mec1-az2, and is still working on the other two.

This is a rare case of a hyperscale cloud region being permanently lost. AWS says most customers restarted operations in other regions by restoring from backups or copying accessible data. Customers whose only copy lived in the affected region lost it. AWS's own guidance is that spreading workloads across Availability Zones does not protect against a region-wide failure; that requires a disaster recovery plan using another region.

Most businesses will never face war damage, but the same lesson applies to fires, floods, power events and provider failures. Ask your provider or IT team where every copy of your critical data lives, including backups, and whether at least one copy is in a different region and ideally a different account. Confirm how long a restore to another region would take and when it was last tested.

The familiar 3-2-1 rule still holds: three copies, on two types of media, with one kept off-site. In the cloud, off-site should mean a separate region.