Microsoft has established the cause of the global outage of its services
Microsoft Corporation has established the causes of the large-scale outage that on November 12–13 disrupted the operation of Outlook, Teams, and other Microsoft 365 services worldwide. The problem arose due to a software bug that appeared back in July 2026, and several technical malfunctions that led to server overload.
The Guardian reported this on November 16, citing preliminary results of Microsoft's investigation and a statement by the company's executive vice president Charles Lamanna.
According to the company, the global outage was the result of a combination of three factors. The first is a hidden software bug in the Microsoft 365 network infrastructure that emerged after an update that began rolling out in July.
The update was in use for more than three months without noticeable problems. However, under a certain load, the bug caused server overload, causing them to stop responding to requests.
Microsoft acknowledged that standard testing did not detect the defect. The test environment used hardware that differed from the servers on which the real services were running. In particular, the problem manifested more strongly on more modern machines with more processor cores.
The second factor was an increase in load on the network infrastructure. Approximately 15 hours before the start of the large-scale outage, the Microsoft Teams system automatically switched to the main network gateway, which also serves Exchange Online.
The third problem was that at that point, about 14% of the server capacity of the relevant infrastructure was already unavailable due to a separate malfunction.
Individually, these circumstances should not have led to a global outage. However, their combination triggered the hidden bug, after which a chain reaction began.
In less than two hours, more than 80% of the network gateway servers stopped responding to requests. Automatic redirection of traffic to the remaining healthy servers only worsened the situation: they also could not withstand the increasing load.
As a result, millions of users worldwide faced problems accessing email and certain features of Microsoft corporate services.
The most serious consequences affected Exchange Online. Users could not access mailboxes through Outlook and other applications, and email delivery was delayed.
As BleepingComputer notes, recovery was particularly difficult due to the accumulated requests. After restarting, servers became overloaded again, and standard automatic recovery mechanisms could not cope with a malfunction of this scale.
In addition, the bug had been introduced several months earlier, so it could not be quickly remedied by a simple rollback of the latest update.
Microsoft engineers had to develop a fix and gradually install it on the damaged infrastructure. The main consequences of the outage were eliminated on November 13, after which the company continued to check the stability of the services.
Microsoft apologized to customers and promised to publish a final technical report. The company also announced an increase in available capacity and additional measures to prevent similar large-scale failures.
At the same time, the identified defect concerned the Microsoft 365 infrastructure and, according to the company's explanations, did not affect ordinary installations of the Windows operating system.