# Microsoft Links 365 Outage to Gateway Bug, Teams Shift and Capacity Loss

By Simon Yoon

Canonical URL: https://www.tokenpost.com/news/technology/58864
Published: 2026-11-16T22:14:36.000Z
Updated: 2026-11-16T22:14:36.000Z
Section: Technology

> Exchange Online was hit hardest, while nearly 14% of gateway server capacity was already offline.

Microsoft linked [last week’s global Microsoft 365 outage](<https://www.tokenpost.com/news/technology/57309>) to a gateway software defect, a Teams traffic shift and lost capacity that overwhelmed the networking gateway serving its services.

The impact window ran from 11:20 a.m. ET Thursday, Nov. 12, through 11:10 a.m. ET Friday, Nov. 13. Exchange Online suffered the greatest impact: most users could not access their mailboxes through Outlook or other email clients, and message delivery was delayed. Teams, Copilot, SharePoint, OneDrive, Defender and Purview were also affected, while most other Microsoft 365 services largely remained online and accessible.

The defect was part of a fix for network gateway servers that began deploying in July. Under certain conditions, lock contention inside a Windows process could exhaust a server’s CPU and memory. Testing tended to focus on weaker hardware with fewer cores, and the traffic mix and hardware models in test labs did not capture the conditions that exposed the defect on newer machines with more cores.

About 15 hours before the impact window began, Teams traffic shifted from an alternate gateway to the primary gateway, which also serves Exchange Online. The change altered traffic patterns and increased load, but the elevated load did not reflect an overall increase in traffic: unavailable servers were shifting demand onto the remaining infrastructure. Nearly 14% of gateway server capacity was already offline because of a separate issue from the previous day. That capacity loss alone had not affected service.

Traffic controls redirected users to healthy endpoints, concentrating demand on fewer servers and helping the failures spread. More than 80% of servers were unresponsive at the low point, reached in less than two hours.

Microsoft restored the connection-handling code path to a previously known good state and returned servers to service in stages. It added capacity and traffic-distribution safeguards and extended the delivery window for queued email. A comprehensive list of repair items is due as part of the final report Friday, Nov. 20.



## Links in this article

- [last week’s global Microsoft 365 outage](https://www.tokenpost.com/news/technology/57309)
