Microsoft confirmed that "elevated load" on its infrastructure caused widespread connectivity failures across Outlook, Teams, and Copilot on Thursday. The company stated that specific servers became unexpectedly unavailable, straining the remaining network components. This issue affected users in multiple regions, including North America and the United Kingdom. Downdetector data showed over 131,000 reports of Outlook problems at 18:15 GMT. A separate spike occurred around 18:30 GMT for UK users.
Infrastructure Failure and User Impact
The monitoring site recorded more than 129,000 incidents related to email services. Microsoft 365 followed with approximately 46,000 reports. OpenAI reported investigating issues with Microsoft connectors where new connections might interrupt tool functionality. Anthropic identified a problem with a top-level provider affecting its Microsoft 365 connector. These disruptions extended beyond basic communication tools into AI platforms. The outage clearly disrupted professional workflows globally. I find it alarming that such a fundamental service failure cascaded into third-party AI ecosystems so quickly. Users could not access their emails or chat applications during the peak hours. The scale of the incident suggests a significant underlying technical flaw rather than a minor glitch.
Resolution Efforts and Technical Details
Microsoft initially resolved a Teams interruption affecting North American users early in the day. Later updates indicated the problem might impact Microsoft 365 connectivity in other regions. At 19:50 GMT, the company announced it was testing a configuration fix. This specific change aimed to address servers failing automated health checks. The firm explained that the elevated load was not driven by increased traffic volume. Instead, the unavailability of certain servers created the bottleneck. Microsoft is working to restore these servers safely. The situation remains fluid as engineers implement the proposed solution. I expect full stability soon, but the initial delay in fixing the core server issue was frustrating. The lack of immediate resolution for global users highlights a gap in redundancy planning. Until the configuration test succeeds, many will remain offline.
