Originally reported November 17, 2026.
A software bug introduced in July contributed to a global Microsoft service outage in November, according to the company’s preliminary post-incident report released on November 15. Outlook, Exchange Online, Teams and Copilot were among the affected services.
Microsoft said the bug could cause network gateway servers to exhaust CPU and memory under particular loads. It went undetected during routine testing and a gradual global rollout. On November 11, a routine move of Teams traffic to the primary gateway increased the load on servers that also supported Exchange Online. Nearly 14% of gateway server capacity was already unavailable because of an unrelated issue the previous day. Microsoft said that capacity loss alone would not have caused an outage.
As servers became unresponsive, automated traffic management shifted requests to other servers exposed to the same bug and load conditions. Microsoft said the failure spread across the relevant server cluster in less than two hours; at its worst, more than 80% of those servers could not respond to requests.
Microsoft identified the severity of the problem at 11:20 a.m. US Eastern time on November 12. After initial recovery attempts failed, it developed an update restoring the affected code to a previously working state. The company said customers were no longer directly affected by 11:10 a.m. Eastern time on November 13—23 hours and 50 minutes after it formally identified the incident.
On November 16, Microsoft executive vice president Charles Lamanna attributed the disruption to the combination of the bug, increased load and lost capacity. The company apologized and said it had increased capacity and strengthened monitoring and traffic-allocation safeguards. Microsoft shares closed that day down 1.12% at $517.99; the original report did not establish that the outage caused the stock move.
Source: m.thepaper.cn (originally published 2026-11-17 14:07 UTC+8)
