A software defect left Microsoft's Outlook email service unavailable to millions of users around the world for nearly 24 hours last week, in what amounted to one of the company's most severe outages. Employees, businesses, and governments lost access to Outlook email as Exchange Online servers, which power the service, went offline starting in Switzerland and then spread to Mexico and other countries. Within roughly two hours, more than 80 percent of Exchange servers were unresponsive, according to a source familiar with the incident.

Microsoft tagged the event as "sev0" in its internal incident tracking system, the highest severity level, signaling a catastrophic disruption. Hundreds of engineers worked to identify the cause and restore service. The company's failover systems moved email traffic toward healthy servers as designed, but that shift also widened the problem. Some servers entered maintenance mode and stopped responding under unusually high CPU loads, and attempts to bring them back online were almost immediately swamped by queued traffic.

Engineers examined crash reports, trace logs, and kernel dumps in the early hours, and Microsoft's normal rollback processes failed to return servers to a healthy state. The company also investigated whether a cyberattack was responsible and consulted DDoS specialists, but crash reports and logs quickly ruled out an attack. Attention then turned to the servers' networking stack, where engineers found an unusual bug that had been present on Microsoft 365 networking gateway servers since July without affecting operations.

Charles Lamanna, who leads Microsoft's Copilot, Agents, and Platform team, said in a post on X that a latent defect caused servers to become overwhelmed and unresponsive under certain load conditions. Microsoft's preliminary post-incident report describes the flaw as affecting select Windows servers due to lock contention inside an internal Windows process. The bug did not affect regular versions of Windows, because it resulted from a fix prepared specifically for Microsoft 365 networking gateway servers, and Microsoft has said its testing labs missed it because they relied on weaker hardware with fewer cores.

Two other incidents combined to make the bug surface. Fifteen hours before the Outlook outage, Microsoft Teams failed over from an alternative networking gateway to the main gateway supporting Exchange Online, increasing load and changing traffic patterns. At the same time, about 14 percent of the gateway's server capacity was out of service because of an unrelated issue from the previous day.

Recovery took longer than usual because servers returning to service were immediately overwhelmed by waiting traffic and became unresponsive again, Lamanna said, adding that Microsoft's automated recovery processes were not built to restore service from such a broad unresponsive state. Engineers maintained a bridge call throughout, at one point reaching its maximum of 1,000 participants, and overlapping interventions from different teams complicated coordination at times. Microsoft is preparing a final post-incident review for publication later this week.

More technology news from TechManNews.