By Hamza Faisal
Microsoft has traced the global Microsoft 365 outage that cut off Outlook and Teams for nearly a full day last week to a Windows bug it introduced in July — a defect that sat dormant for more than three months before a surge in traffic brought down more than 80 percent of the servers that front its cloud services.
The company's preliminary post-incident report, dated November 15 and shared with administrators, says the outage ran from 4:20 p.m. UTC on November 12 to 4:10 p.m. UTC on November 13. Exchange Online, the email system behind Outlook, was hit hardest: most users could not reach their mailboxes and delivery was delayed for hours. Teams, OneDrive, SharePoint and Copilot were also affected.
A July fix that quietly waited three months
The faulty change was written specifically for the networking gateway servers that sit in front of Microsoft 365, receiving users' connections before passing them on to services such as Exchange. Under certain loads, it caused tasks inside an internal Windows process to get stuck waiting on one another — a problem known as lock contention — until the machines ran out of processing power and memory and stopped responding.
The change went through Microsoft's staged deployment from July and ran worldwide for more than three months without visible impact. Microsoft said its testing missed the defect because its laboratories mainly used less powerful hardware with fewer processor cores. The contention gets worse on newer machines with more cores.
Microsoft stressed that the defect and the fix that caused it were specific to its Microsoft 365 gateway infrastructure and do not affect other Windows systems or ordinary desktop PCs.
Three failures lined up at once
The report names three factors that combined to turn a dormant bug into a global outage. First, the July defect itself. Second, about 15 hours before the outage, Microsoft's systems automatically moved Teams traffic from a second, independent gateway onto the one that also carries Exchange Online — a normal part of operations, but one that piled extra load onto the affected servers. Third, nearly 14 percent of that gateway's capacity was already out of service because of an unrelated problem the day before.
With the same amount of load spread across fewer servers, the fault cascaded from server to server. Within two hours, more than 80 percent of the gateway's servers had stopped responding, and the barriers meant to stop a failure spreading between regions did not contain it.
Teams, OneDrive and SharePoint came back earlier because they could switch to the second gateway. Email could not be moved to Microsoft's backup gateway in the same way, which is why Outlook stayed down longest.
What Microsoft did — and what comes next
Engineers applied an update that returned the connection-handling code to a previously known-good state, rolling it out to groups of servers before restarting them. Microsoft began applying that fix about 15 hours after the first alert went out.
The company said the outage was a technical failure with no indication of malicious activity, and it was not related to a cyberattack or to AI agents — a point it had made while the outage was still unfolding.
Microsoft plans to publish a final post-incident report, with a full list of repairs, by November 20. The incident was the fourth major Microsoft 365 outage this year tied directly to the company's own changes and maintenance, raising fresh questions about how a fix can pass staged deployment, run silently for months, and still take down a service used by businesses and governments worldwide.



