Microsoft Traces Nov. 12 and 13 Microsoft 365 Outage to a Defect in a July Gateway Fix
Microsoft says a defect in a July fix for its Microsoft 365 gateway servers caused a global outage on Nov. 12 and 13, with Exchange Online hit hardest.

Microsoft 365 services, including Outlook email and Microsoft Teams, were disrupted from Nov. 12 to Nov. 13, 2026. In a preliminary post-incident report dated Nov. 15, Microsoft attributed the outage to a defect in a fix for its Microsoft 365 network gateway servers, according to Cybersecurity News.
WBAY reported on Nov. 13 that Microsoft described the disruption as a technical issue, not a cybersecurity breach. The Register reported on Nov. 16 that the final post-incident report is due on Nov. 20.
A July Fix That Failed Under Load
According to the preliminary report as summarized by Cybersecurity News, a fix for Microsoft 365 networking gateway servers was deployed starting in July. It ran globally for more than three months without visible problems. The fix contained a hidden defect in an internal Windows process, in which competing tasks waited on shared resources. Under certain traffic conditions, that lock contention exhausted processor and memory capacity.
Two conditions exposed the defect, the report says. About 15 hours before the outage, Teams traffic moved to a primary gateway shared with Exchange Online. Separately, nearly 14% of gateway capacity was unavailable because of an unrelated issue. Automated recovery then moved connections to healthy endpoints, which spread the fault across regions.
The Register reports that Microsoft’s load testing tended to focus on weaker hardware with fewer cores. Microsoft says the defect does not apply to general Windows deployments, according to Cybersecurity News.
WBAY describes the change as introduced during a network capacity upgrade. That differs from the July fix described by The Register and Cybersecurity News.
Preliminary Report Reference
Cybersecurity News reports that Microsoft’s preliminary post-incident report is dated Nov. 15, 2026, and cites Microsoft service health records EX1550543 and MO1550821.
Two Accounts of the Outage Timeline
The two accounts time different milestones. Cybersecurity News puts the outage window at 4:20 p.m. UTC on Nov. 12 to 4:10 p.m. UTC on Nov. 13. The Register gives a Teams failover at 01:07 UTC on Nov. 12, a detected connectivity anomaly at 16:03 UTC that day, a mitigation at 07:20 UTC on Nov. 13, and confirmed health at 16:10 UTC on Nov. 13.
The Register says Microsoft’s final customer update came at 04:15 on Nov. 14. Both accounts describe a similar recovery. Engineers returned Windows connection handling to an earlier working state, removed large server groups from service, applied the change, and restored traffic gradually.
Exchange Online Hit Hardest
Exchange Online was the most affected service. The Register reports that at its lowest point more than 80% of servers were unresponsive. Cybersecurity News reports that most users could not reach mailboxes through Outlook or other clients, and that many experienced delayed email.
Teams, SharePoint, OneDrive and Office apps could shift eligible traffic away from the faulty infrastructure. Exchange Online cannot fail out of the network gateway for scale reasons, according to The Register.
WBAY reports that Outlook, Teams and Copilot were inoperable for millions of users. It cites Downdetector, a third-party tracker, for an outage that affected 74 countries and territories. WBAY also reports that the total number of people affected has not been shared.
WBAY reports that the company reportedly has more than 450 million paid Microsoft 365 commercial subscriptions, but it does not say how many of those were affected.
Microsoft’s preliminary report points to a July fix, and its final report is due on Nov. 20.
Repair Timeline and Customer Guidance
Microsoft applied the repair to groups of servers and gradually restored service, The Register reports. Microsoft says it will publish a full post-incident report once its review is complete, according to WBAY. The Register reports that the final report is due on Nov. 20.
Microsoft’s guidance, as reported by Cybersecurity News, says no tenant or device configuration changes are needed. The company also extended the email delivery window so that queued messages were not returned as undeliverable, and added capacity.
For businesses that run email and collaboration on Microsoft 365, the record as of Nov. 16 covers the timeline, the preliminary cause as Microsoft describes it, and the guidance above. The final report is due on Nov. 20.
Get more stories like this in your inbox by signing up for the EW newsletter, The Entrepreneurs Weekly’s free digest of founder stories and business news.





