Microsoft says a software flaw from July increased network demand and reduced server capacity combined to trigger last week’s widespread Microsoft 365 disruption.
Microsoft has apologised to customers after detailing the sequence of events that led to its major Microsoft 365 outage on November 12.
Charles Lamanna, Microsoft’s Executive Vice President of Copilot, Agents and Platform, shared a preliminary incident report on X explaining how a dormant software bug, heavier network traffic and reduced server availability combined to cause a worldwide failure of networking infrastructure used by several Microsoft 365 services.
A statement read: “We are deeply sorry for the impact to all our customers and recognize the impact that this has on their normal business operations.”
According to Microsoft, the outage centred on a networking gateway that sits in front of many Microsoft 365 services. Engineers discovered a hidden defect that could cause servers to become overwhelmed and stop responding under certain traffic conditions.
The company said the issue stemmed from a gateway-specific software fix that began rolling out in July. The change passed Microsoft’s usual deployment processes and ran globally for more than three months without triggering any problems.
However, Microsoft admitted its testing environment failed to uncover the fault because it relied on a different mix of traffic and lower-powered hardware than the systems operating in production.
A second factor emerged around 15 hours before the outage, when Microsoft Teams traffic was shifted onto a primary networking gateway that also supports Exchange Online. While the transfer was described as part of routine resilience procedures, it increased demand on the affected infrastructure.
At the same time, roughly 14 per cent of the gateway’s server capacity was unavailable because of an unrelated issue from the previous day.
Microsoft said neither event would have caused disruption on its own, but together they exposed the dormant bug, causing requests to accumulate until servers began failing during a period of peak US traffic.
Automated systems attempted to reroute users to healthy endpoints, but this unintentionally spread the problem across the remaining infrastructure.
Microsoft explained: “This caused a cascading failure from server to server as the same amount of load was spread across fewer servers.”
Within two hours, more than 80 per cent of the affected servers had become unresponsive, allowing the impact to spread globally after separate safeguards designed to contain such failures did not perform as intended.
Services including Teams, OneDrive, SharePoint and Office applications were able to move away from the faulty gateway infrastructure, enabling them to recover sooner. Exchange Online, however, remained dependent on the affected system.
Recovery efforts were also complicated by the scale of the incident, with returning servers immediately becoming overwhelmed by queued traffic.
Once engineers identified the root cause, Microsoft developed a corrective update and gradually restored service by taking groups of servers offline, applying the fix and bringing them back online in stages.
The post added: “More details on the root cause can be found in the Microsoft 365 Admin center, and an abridged version can be found below.”
Microsoft said a full Post-Incident Review containing additional technical details will be released later this week.
Microsoft reveals three-factor chain behind global Microsoft 365 outage







