AWS Data Centre in UAE Hit in Iranian Strikes
As hostility between the US and Iran continues, Amazon Web Services (AWS) confirms that one of its data centre facilities in the UAE has been caught in the crossfire, triggering a fire and service disruption across part of its ME-CENTRAL-1 region.
AWS states that unidentified objects hit an Availability Zone in the early hours of Sunday 1 March. The incident is widely speculated to link to Iranian strikes on the UAE and other Gulf Arab states, in response to US-Israeli action that killed Supreme Leader Ayatollah Ali Khamenei.
The most recent update on the AWS Health dashboard (2 March, 6:22AM PST) reads: “We continue to work towards recovery of the two impaired Availability Zones (mec1-az2 and mec1-az3) in the ME-CENTRAL-1 Region.
“We are expecting recovery to take at least a day, as it requires repair of facilities, cooling and power systems, coordination with local authorities, and careful assessment to ensure the safety of our operators.
“EC2, Amazon DynamoDB and other AWS Services continue to experience significant error rates and elevated latencies.
“We recommend customers enact their disaster recovery plans and recover from remote backups into alternate AWS Regions, ideally in Europe. Further, we strongly advise customers to update their applications to ingest S3 data to an alternate AWS Region.”
Availability Zones under pressure
Availability Zones, or AZs, are designed as discrete data centre facilities within a single geographic region. Each zone has independent power, cooling and networking, enabling customers to distribute workloads across multiple sites to reduce the risk of single-point failure.
In the UAE, mec1-az2 is one of the primary zones serving local and regional customers, hosting compute, storage and database resources. After the fire, AWS continues to address what it describes as a localised power issue affecting multiple AZs in ME-CENTRAL-1, specifically mec1-az2 and mec1-az3.
The affected zones experience increased EC2 API errors and failed instance launches. Amazon Elastic Compute Cloud, or EC2, allows customers to create and manage virtual servers. Failed launches mean organisations cannot provision new capacity in those zones and may struggle to manage existing workloads.
Key platform services also report disruption. Amazon DynamoDB, a managed NoSQL database service, and Amazon S3, the company’s object storage platform, show delays and elevated error rates. S3 is architected to tolerate the loss of a single AZ, yet the impairment of more than one zone within the region places pressure on service performance.
Existing resources in mec1-az1 remain operational, underscoring the value of distributing applications across zones. Some financial institutions and banks in the UAE report online banking platform disruptions, indicating that the impact extends beyond internal AWS operations to end users dependent on cloud-hosted infrastructure.
Power, redundancy and recovery
According to the AWS update, the fire department has shut off power to the impacted facility and its generators to contain the blaze. Only after clearance can power restoration begin. Data centre recovery therefore depends not only on internal engineering response but also on external safety authorities.
AWS has advised customers to back up critical data and, where possible, fail over workloads to another AWS region. A region consists of multiple AZs in a specific geographic area. Multi-region architecture allows applications to shift traffic to a different country or continent if an entire region becomes impaired.
The AWS management console and command line interface also experienced disruption, limiting customer ability to manage resources across the impacted zones. Engineers are conducting assessments of data health within affected storage systems and state that any required repairs will take place before full restoration.
It is not the first time that AWS data centres have experienced global service outages – on 20 October, AWS faced disruption across core data centre services DynamoDB and EC2.
The most recent incident places focus on established data centre design principles, including geographic separation and workload distribution across multiple AZs and regions. It also demonstrates the limits of resilience when physical infrastructure faces direct external impact.
The events in the UAE underline the operational realities of running cloud infrastructure in geopolitically sensitive regions, where resilience planning extends beyond technical architecture to encompass location strategy and cross-border redundancy.
