Amazon Web Services Reports Hours-Long Recovery Effort After UAE Data Center Power Failure

On March 1–2, 2026, Amazon Web Services (AWS)  reported a significant operational disruption affecting multiple Availability Zones in the ME-CENTRAL-1 Region, specifically mec1-az1, mec1-az2, and mec1-az3. The issue originated from a localized power failure in mec1-az2, caused by an object striking the data center and sparking a fire, which required intervention by the fire department. As a result, power was shut off to the facility and generators, leading to widespread service impacts.

Customers experienced increased error rates and latencies across multiple AWS services, including EC2, DynamoDB, and S3. EC2 API errors and instance launch failures were reported, and new instance launches were largely unavailable in the affected zones. Some network-related APIs, such as AllocateAddress, AssociateAddress, DescribeRouteTable, and DescribeNetworkInterfaces, were also impacted. While existing resources in unaffected AZs continued operating normally, customers with workloads in the affected zones faced significant disruptions.

AWS implemented mitigations to support recovery, including enabling the disassociation of Elastic IP addresses from affected resources and redirecting traffic to unaffected AZs. Customers were advised to failover to alternate Availability Zones or AWS Regions and to back up critical data. Over the course of the incident, AWS reported gradual improvements for certain API operations, but full recovery of the impacted resources remained several hours away.

Updates were provided periodically, highlighting continued mitigation efforts, ongoing recovery of EC2 networking APIs, and restoration steps for other services. AWS emphasized that customers using redundant deployment across multiple AZs were not impacted by the outage. The company continues to work on restoring power, connectivity, and normal service operations across the affected zones.