The Microsoft Azure OpenAI Service experienced a major outage in the Sweden Central region, impacting users throughout the working day. The issues began early in the morning, with Microsoft acknowledging the problems at 0900 UTC and indicating that they were due to an “unhealthy backend dependent service” that triggered cascading failures.
Service Disruption Timeline
Users encountered errors while attempting to access various models, including GPT-5.2, GPT-5 Mini, and GPT-4.1. Microsoft attempted to mitigate the situation by rebooting the affected service at 1236 UTC. However, the problems persisted, with reports of pods crashing due to out-of-memory errors at 1246 UTC. In response, Microsoft began scaling out nodes in the cluster to enhance request handling and resilience.
Resolution Efforts
By 1530 UTC, Microsoft had increased the memory available to the pods, completing this adjustment by 1553 UTC. The service was confirmed to be operational again at 1612 UTC, coinciding with the end of the workday for many users in Sweden.
Community Reactions and Implications
While Microsoft’s transparency in acknowledging the outage received praise, the duration of the disruption raised concerns about the reliability of its cloud services. Comments on social media reflected a mix of humor and frustration, with one user noting, “EU resilience is getting another live exercise.” Others viewed the incident as a learning opportunity, emphasizing the importance of building resilience into cloud infrastructure.
Impact on Azure’s Reputation
This incident poses potential reputational risks for Microsoft as it seeks to encourage broader adoption of its AI services. The lengthy recovery time from what appears to be a software issue may deter customers from fully leveraging Azure’s capabilities, particularly in critical applications.
This article was produced by NeonPulse.today using human and AI-assisted editorial processes, based on publicly available information. Content may be edited for clarity and style.
