AWS Power Outage in Middle East Triggers Major Disruption to EC2 and Networking Services
On Sun, Mar 1, 2026, a power outage occurred in the AWS me-central-1 region due to an incident involving external objects striking a data center. This led to sparks and a subsequent fire.
On Sun, Mar 1, 2026, a power outage occurred in the AWS me-central-1 region due to an incident involving external objects striking a data center. This led to sparks and a subsequent fire.
This incident affected Amazon Elastic Compute Cloud (EC2) services, networking APIs, and resource availability in a single Availability Zone (mec1-az2).
The local fire department required a complete shutdown of the power supply to the facility, including backup generators, to manage the situation. This resulted in the incapacitation of EC2 Instances, Amazon Elastic Block Store (EBS) volumes, and Amazon Relational Database Service (RDS) databases within the affected zone.
The disruption started around 4:30 AM PST, with AWS investigating connectivity and power issues by 4:51 AM PST. By 6:09 AM PST, AWS confirmed the localized power failure in mec1-az2.
Traffic weighting strategies were implemented to route requests away from the damaged facility to other Availability Zones within the region.
The outage significantly impacted EC2 networking APIs, with widespread throttling errors and failures in critical networking functions such as AllocateAddress, AssociateAddress, DescribeRouteTable, and DescribeNetworkInterfaces.
On Sun, Mar 1, 2026, a power outage occurred in the AWS me-central-1 region due to an incident involving external objects striking a data center.
Throughout the afternoon, AWS applied multiple configuration changes to address API failures. By 2:28 PM PST, the AllocateAddress API began showing signs of recovery.
At 6:01 PM PST, AWS confirmed the recovery of the AssociateAddress API requests. An update was deployed to allow customers to disassociate Elastic IP addresses from resources in the affected data center.
This enabled connectivity restoration by associating existing IP addresses with new resources in unaffected zones.
Despite API functionality restoration, physical infrastructure remained offline, pending clearance from local authorities for power restoration.
AWS emphasized the significance of deploying applications across multiple Availability Zones, which largely insulated such setups from the outage.
Organizations needing immediate recovery were advised to launch replacement resources in unaffected zones or within other AWS Regions, utilizing recent EBS snapshots or backups.
Owing to the traffic shift from the affected zone, customers might experience longer provisioning times or require retries when launching specific instance types in healthy ME-CENTRAL-1 zones.
As of the last update at 6:01 PM PST, there was no estimated time for power restoration at the mec1-az2 facility. AWS continues to recommend operating out of alternate Availability Zones or Regions as recovery efforts proceed.
Based on reporting by Cyber Security News.
