A significant service disruption struck the Middle East infrastructure of Amazon Web Services after objects hit a data center facility in the UAE, sparking a fire and forcing a power shutdown in one of its Availability Zones.
According to AWS status communications issued March 1, the incident began around 4:30 AM PST when debris struck the mec1-az2 facility inside the ME CENTRAL 1 Region. The impact caused sparks and ignited a fire. Emergency crews then cut power to the building and its generators while they contained the situation. As a result, core services in that zone went offline.
The outage initially affected Amazon Elastic Compute Cloud instances, Elastic Block Store volumes, and Relational Database Service databases. Immediately after the incident, AWS reported that another Availability Zone, mec1-az3, also suffered a localized power issue. Meanwhile, customers operating in mec1-az1 reported elevated API error rates and difficulties launching new EC2 instances. At one point, AWS stated that customers could not launch new compute instances within the region, although existing workloads in unaffected zones continued running.
Beyond compute and storage, the disruption cascaded across a broad range of services. AWS listed degraded performance in DynamoDB, Cognito, Lambda, S3, Elastic Kubernetes Service, Redshift, and CloudWatch, among dozens of others. In response, AWS advised customers to shift workloads to alternate Availability Zones or fail over to different geographic regions where architectures allowed.
Separately, AWS acknowledged increased error rates in the ME SOUTH 1 Bahrain Region due to a localized power issue in a single zone there. Engineers began routing traffic away from the affected infrastructure to stabilize performance.
While AWS attributed the primary outage to objects striking the facility, the company has not disclosed further detail about the nature of those objects or whether external factors played a role. The incident happens at a time of heightened tensions, yet AWS has not identified the cause of the incident other than the physical impact and the resulting fire.
For most organizations, the incident highlights the value of a redundancy strategy based on multiple zones and multiple regions. For some users, resilient architectures have minimized the impact of the incident, but for others, the incident has meant hours of degraded performance while AWS works to bring power back online.
