Microsoft has resolved the worldwide Microsoft 365 outage that disrupted Outlook, Teams, Copilot, and other services on November 12–13. In a preliminary incident report, the company traced the nearly 24-hour disruption to a months-old Windows Server defect, increased network traffic, and reduced server capacity.
The outage began at 4:20 p.m. UTC on November 12 and ended at 4:10 p.m. UTC the following day. Exchange Online was hit hardest: many users could not access their mailboxes or send and receive email normally. Teams, Copilot, SharePoint, and OneDrive also experienced issues with features that rely on Exchange Online.
A dormant Windows bug set off a cascading failure
Microsoft identified three factors behind the disruption. The first was a defect introduced in a Windows update deployed to Microsoft 365 networking gateway servers in July 2026.
Under certain workloads, the defect caused an internal Windows process to experience lock contention. CPU and memory use rose until affected servers became unresponsive. The issue went undetected for more than three months because Microsoft’s test environments primarily used less powerful hardware with fewer CPU cores, where the problem was less likely to appear.
The second factor was a change in Teams traffic. About 15 hours before the outage, traffic was automatically redirected from an alternative gateway to the primary gateway shared with Exchange Online. Although such failovers are routine, the move added load to the affected infrastructure.
At the same time, around 14% of gateway server capacity was already unavailable because of an unrelated issue from the previous day.
As servers stopped responding, automated traffic-management systems shifted requests to the remaining machines, adding further pressure. Within two hours, more than 80% of the networking gateway servers had become unresponsive. Safeguards intended to limit failures to individual regions did not prevent the disruption from spreading worldwide.
Microsoft said the Windows defect was specific to its Microsoft 365 gateway infrastructure and does not affect Windows deployments more broadly.
Recovery and additional safeguards
Restoring service proved difficult. Servers often became unresponsive again after restarting as accumulated traffic returned, while standard rollback procedures were ineffective because the defective change had been deployed months earlier.
Engineers developed a fix that restored the affected connection-handling code to its previous working state. Microsoft began returning infrastructure to service on November 13, and webmail availability was back to normal worldwide by 2:17 p.m. UTC. The company later confirmed the incident was resolved through its Microsoft 365 Status account, with final service-health notifications issued on November 14.
Microsoft has added traffic-distribution safeguards, expanded infrastructure monitoring, and increased capacity. During the incident, it also extended email-delivery timeouts to protect queued messages and confirmed that delayed mail had since been processed.
The company said there was no indication that the outage involved malicious activity or AI agents. Its investigation is ongoing, and Microsoft plans to publish a final incident report by November 20, 2026.
