Amazon Web Services (AWS) is recommending that customers affected by the outage at its data centers in the United Arab Emirates and Bahrain move their workloads to other cloud regions, more than six months after the damage suffered during the conflict between the United States, Israel, and Iran. AWS has confirmed it will not be able to recover resources and data hosted exclusively in one of its UAE availability zones, while it continues working on the rest of the affected infrastructure.
The key facts about AWS’s Middle East data centers in 20 seconds
- AWS will not be able to restore resources hosted exclusively in the UAE’s mec1-az2 zone.
- The company continues working on the other affected zones in the UAE and in Bahrain.
- AWS recommends migrating applications to other regions and recovering data from external backups when necessary.
- The company still hasn’t given a specific date for fully restoring services.
- The UAE is exploring, according to Tom’s Hardware, alternatives such as underground data centers to address security risks.
The situation is one of the longest infrastructure outages at a major cloud provider following physical damage to data centers. The problems began in March, when several AWS facilities in the region were damaged during attacks tied to the conflict, an episode CloudNews covered as it unfolded. In the UAE, AWS confirmed that two facilities were hit directly, while a facility in Bahrain suffered damage from a nearby explosion.
Six months later, the recovery is still incomplete. AWS has updated its status information to warn that one of the availability zones in its UAE region will not be able to recover the resources and data that remained hosted there exclusively. The company is continuing recovery work on the other affected zones and says it is replacing part of the damaged infrastructure.
One of AWS’s UAE zones can no longer be recovered
The Middle East (UAE) region, identified as me-central-1, is made up of three availability zones: mec1-az1, mec1-az2, and mec1-az3. AWS uses these physically separate zones to distribute resources and improve the resilience of its services. Bahrain has its own region, me-south-1, also split into three zones.
After the damage in March, AWS began recommending that customers activate their disaster recovery plans and move any workloads that were still accessible to other regions. In a later update, the company explained that the recovery process in the UAE would take several months.
The diagnosis is now more specific. AWS says that, after assessing the damage, it cannot restore access to the resources and data hosted exclusively in mec1-az2. The company continues working on mec1-az1 and mec1-az3 and says it is replacing the affected infrastructure.
For customers, the difference matters. An application designed to spread its resources across multiple regions, or one with external backups, can rebuild its service elsewhere. Data that existed only in the affected zone, however, cannot be recovered through a simple migration.
AWS says most customers have already been able to restore operations in other regions using backups or by copying data that remained accessible. The company is also continuing to offer support to help move applications to alternative regions.
Bahrain still hasn’t fully recovered either
The situation also remains unresolved in ME-SOUTH-1, AWS’s region in Bahrain. Reuters reported on September 15 that AWS could not recover access to the affected infrastructure in Bahrain and to one of the UAE’s three availability zones.
In the early days of the crisis, AWS explained that a facility in Bahrain had been hit by a nearby explosion and that teams were working to restore power and connectivity. The company recommended at the time that customers migrate or replicate their data to other regions.
The damage in the region affected companies that relied on AWS infrastructure for various services. Reports published during the first months of the outage pointed to problems for financial and technology companies in the region, as AWS tried to move customer workloads to other locations.
AWS has not given a firm timeline for full recovery. In its most recent update on the UAE, the company says it will provide further information over the coming months. For Bahrain, the communications cited by Tom’s Hardware also don’t set a specific recovery date.
The recommendation to migrate doesn’t mean that all of AWS’s Middle East services are down. The problem affects specific infrastructure and availability zones, and AWS continues operating resources that weren’t affected or that have already been recovered.
The impact goes beyond repairing servers
The physical damage suffered by the data centers shows a different kind of limitation than the ones cloud infrastructure usually faces. Redundancy between availability zones is designed to protect services against individual failures, but several concentrated physical incidents within the same region can exceed those safeguards.
AWS explains that its availability zones are physically separated and have independent power, cooling, and connectivity resources. That architecture is designed to let applications keep running even if one zone becomes unavailable, the same resilience model AWS has kept pointing to throughout the months of recovery in the Gulf.
The Gulf episode has also put the physical location of data centers on the table. Tom’s Hardware reports that UAE authorities have raised the possibility of developing underground data centers as one of the alternatives to protect future facilities against drone attack risks. The use of dedicated defense systems to protect these complexes has also been discussed.
These measures would come at a high cost and are not part of any solution AWS has announced for its damaged facilities. They are options mentioned in the broader debate about how to protect the digital infrastructure the UAE wants to keep developing.
The country has spent years trying to attract investment tied to data centers, artificial intelligence, and cloud services. The damage suffered by AWS adds a new variable to that development: beyond electricity, connectivity, and land availability, new facilities now have to weigh the physical and geopolitical risks of their location.
The most immediate consequence for AWS customers is more practical. Applications that depended on a single region or availability zone need proper recovery mechanisms in place. Backups located in another region, data replication, and multi-region architectures can all reduce the impact of a prolonged physical outage.
The case also shows that the word “cloud” doesn’t remove the dependency on physical facilities. Digital services still need buildings, electrical systems, cooling, networks, and equipment that can be damaged by external events. In this case, repairing part of that infrastructure has dragged on since March, and AWS still hasn’t set a firm date for completing the recovery.
Frequently asked questions
What happened to AWS’s data centers in the UAE?
Several facilities were damaged during the attacks in March. AWS continues to recover part of the infrastructure, but it has confirmed it cannot restore the resources and data hosted exclusively in the mec1-az2 zone.
Does AWS recommend migrating data to another region?
Yes. AWS recommends that affected customers move their applications to alternative regions and use backups or accessible data to rebuild services when necessary.
When will AWS be fully operational again in the UAE and Bahrain?
AWS has not given a specific date. In the UAE, it says it is replacing damaged infrastructure and will provide updates over the coming months.
What is an AWS availability zone?
It’s physically separate infrastructure within an AWS region, designed with independent power, cooling, and connectivity resources. The UAE and Bahrain regions each have three availability zones.
