Resolved

Recovery has been completed, and the incident has been fully resolved.
Service was progressively restored through controlled recovery procedures following successful infrastructure isolation, validation, and capacity rebalancing. Extended health verification has confirmed stable AI model serving operations across the affected region.

We appreciate your patience while our Engineering, Infrastructure, and Site Reliability teams executed the recovery. A comprehensive post-incident review is now underway to identify further opportunities to strengthen operational resilience.

Monitoring

Mitigation has been successfully completed, and service has been restored across the affected AI Model Serving infrastructure.

We are currently performing controlled validation and continuously monitoring network health to verify service stability, workload distribution, and operational integrity before declaring the incident fully resolved.

The affected GPU serving infrastructure has been successfully isolated to contain the impact and prevent further service degradation.

Traffic has been redirected to healthy serving capacity, and our engineering teams continue executing recovery procedures while validating AI model serving operations.

As a precaution, we have activated additional traffic safeguarding and infrastructure protection measures across our global AI serving platform while recovery efforts continue.

We have initiated mitigation and recovery procedures as our on-call engineers continue working to restore affected services. Recovery efforts are progressing, and service availability is gradually improving across impacted infrastructure.

Identified
We have identified the issue affecting our AI Model Serving service in the Bahrain region. The disruption is due to a regional infrastructure impact associated with an ongoing political event, affecting a portion of our distributed GPU capacity in the region.

Some AI models may experience increased latency, intermittent errors, or temporary unavailability while recovery efforts are underway. Our engineering team is actively mitigating the impact and working to restore full service as quickly as possible.

We will continue to provide updates as more information becomes available.

We are aware of an ongoing issue impacting our distributed GPU infrastructure in the Bahrain region. As a result, some AI models may experience service disruptions, increased latency, or temporary unavailability.