Microsoft has identified a software defect, increased network traffic, and reduced server capacity as the three factors behind last Thursday’s global Microsoft 365 outage, which disrupted Outlook, Exchange Online and several other services.

Charles Lamanna, Microsoft's executive vice president of Copilot, Agents and Platform, outlined the cause in a post on X, apologizing to customers for the disruption to their businesses and daily operations.

The underlying software defect was introduced in an update that began rolling out in July. Under certain traffic conditions, the update caused servers to become unresponsive. Although the software operated worldwide for more than three months without incident, Microsoft's testing procedures failed to identify the problem.

According to Lamanna, the defect was more pronounced on newer machines with greater numbers of processor cores, while the company's load testing focused primarily on older, less powerful hardware.

Two additional circumstances contributed to the failure. Approximately 15 hours before the outage, Microsoft Teams switched from a backup network gateway to the primary gateway, which also supports Exchange Online. The change was part of Microsoft's normal operations, but it increased the amount of traffic handled by the gateway.

At the same time, nearly 14% of the gateway's servers were already out of service because of an unrelated problem from the previous day. The combination of increased traffic and reduced capacity triggered the software defect during peak U.S. usage hours.

Microsoft's systems initially attempted to redirect users to functioning servers, but the additional traffic caused the same problem to spread. Within two hours, more than 80% of the gateway's servers had become unresponsive. A separate issue also prevented safeguards designed to isolate disruptions between geographic regions from containing the failure.

Outlook and Exchange Online experienced the most significant disruptions, while problems also affected Teams, Microsoft 365 Copilot, SharePoint Online, Intune and Defender for Office 365. Reports of Outlook problems on Downdetector peaked at approximately 130,000.

Teams, OneDrive, SharePoint and Office applications recovered sooner because they were able to switch to other network gateways. However, restoring the affected servers proved more difficult than usual. Servers returning to operation were overwhelmed by accumulated traffic, causing some to fail again.

Microsoft also could not rely on its standard rollback procedures because the defective update had been in place for several months. Engineers instead developed a new update that restored the software to its previous working state. They deployed the fix across large groups of servers, restarting the machines before returning them to service.

Lamanna said Microsoft plans to release a final post-incident review later this week.