AWS Middle East outage affects UAE and Bahrain amid Iran strikes

Amazon Web Services is dealing with an AWS Middle East outage after an availability zone in its UAE region was “impacted by objects” that caused sparks and a fire, prompting local responders to shut down power to parts of the facility, according to reporting. AWS also flagged separate connectivity and power issues affecting an availability zone in Bahrain, compounding disruption for customers running single-region workloads.

What happened during the AWS Middle East outage

AWS’s public service updates indicated that one UAE availability zone was taken offline after the incident, while other zones in the region continued operating. The company said recovery could take “many hours”, and advised customers to fail over to other availability zones or other regions where their architectures allow. Multiple AWS services were affected, which is typical when a single zone’s power and network connectivity are disrupted.

Reuters linked the timing to the same day Iran launched missile and drone attacks across Gulf states, described as retaliation for earlier US and Israeli strikes. AWS did not publicly confirm the origin of the “objects” and, in Reuters’s account, declined to say whether there was a direct connection to Iranian strikes. Even so, the AWS Middle East outage shows that cloud risk in conflict-adjacent regions is not only about cyberattacks and component failures, but also about physical access, airspace disruption, emergency response decisions, and the realities of local power shutdowns.

What it means for cloud resilience planning

For organisations using the UAE and Bahrain regions to meet latency or data residency requirements, the main lesson is the unglamorous one: if your application is not designed for multi-AZ operation, an availability zone failure becomes a full service outage. Cross-zone replication, tested failover, and clear “degraded mode” behaviour matter more than ever when the failure mode is abrupt and external. As previously reported by eeNews Europe when a data centre fire highlighted infrastructure fragility, power events tend to cascade quickly from facilities into supply chains and customer operations.

In practical terms, this kind of outage is the scenario that validates (or exposes) your assumptions: whether backups are truly independent, whether failover can be triggered without manual intervention, and whether your “regional” design actually survives the loss of a whole availability zone.