AWS reported connectivity and power issues in Bahrain Availability Zone mes1-az2, with impacts to instances, APIs, and other service workflows.
AWS Health Dashboard
health.aws.amazon.com
Production Incident Atlas
AWS Bahrain outage (2026): regional infrastructure unavailable
A facility in AWS’s Bahrain region suffered physical impact and power disruption. AWS later said the region was unavailable and it could not restore resources and data that had not been migrated.
AWS Middle East (Bahrain) Region (ME-SOUTH-1)Software & infrastructureOther failure modeCritical severityAvailability
What happened
Incident timeline
AWS said it had shifted traffic away from the affected zone for most services. Power had not been restored, and affected EC2, database, and storage resources needed a longer recovery.
AWS Health Dashboard
health.aws.amazon.com
AWS said a drone strike near a Bahrain facility caused physical impacts to infrastructure, including disrupted power and, in some cases, water damage from fire suppression.
AWS Health Dashboard
health.aws.amazon.com
Amazon publicly said the Bahrain region had been disrupted by the ongoing conflict and that it was helping customers migrate applications to alternate AWS regions.
Amazon Web Services
aboutamazon.com
AWS said the Bahrain region was unavailable, advised customers to recover resources in other regions from remote backups, and suspended relevant billing while working toward restoration.
AWS Health Dashboard
health.aws.amazon.com
AWS said damage spanned multiple Availability Zones and exceeded its regional and multi-zone design limits. It said it had exhausted options to restore resources and data that had not been migrated before the region became unavailable.
AWS Health Dashboard
health.aws.amazon.com
Detection
AWS began investigating connectivity and power issues in the Bahrain region at 9:56 PM PST on March 1. At 11:09 PM, it reported a localized power issue in mes1-az2 and began routing requests away from that zone.
Recovery
AWS shifted some service traffic away from the affected zone and made software mitigations, but power and resource recovery required longer work. On April 30, AWS said the Bahrain region was unavailable. On September 15, it said it could not restore access to resources and data that had not been migrated before the region became unavailable.
Lessons afterward
AWS recommended using another Availability Zone or region, restoring from remote backups, and migrating workloads. AWS later said this physical damage spanned multiple Availability Zones and exceeded what its regional and multi-zone services were designed to withstand. A separate-region recovery plan was therefore necessary for resources that needed to survive regional unavailability.
Editorial analysis
What happened, in plain language
AWS reported a localized power issue in one Bahrain Availability Zone on March 1. In an update the next day, it said a drone strike near one of its facilities had caused physical impacts to infrastructure. AWS shifted some service traffic away from the affected zone while warning that EC2 instances, database instances, and storage resources in that zone would take longer to recover.
On April 30, AWS said the Bahrain region was unavailable and advised customers to recover resources in other regions from remote backups. On September 15, AWS said the damage spanned multiple Availability Zones and exceeded what its regional and multi-zone services were designed to withstand. AWS said it had exhausted options to restore resources and data that had not been migrated before the region became unavailable.
The practical lesson is that availability zones within one region are not a substitute for a separate-region recovery plan. That is an editorial takeaway from the documented event; AWS specifically recommended migration and remote backups. The status below is based on AWS’s September 15 update, which said another update would come in early 2027.
Editorial severity: Critical, based on AWS’s report that the region became unavailable and some unmigrated resources could not be restored. AWS did not assign this editorial label.
Last reviewed 10/9/2026 by Nexus Lead Architect.
Change notes
- 10/9/2026: Initial publication using AWS public status reports and official Amazon updates.
Corrections or additional primary evidence? Contact the editorial team.