The global disruption of Microsoft 365 that lasted more than 30 hours was traced to a technical flaw introduced during a past network capacity upgrade.
On Nov. 16, Microsoft said in an incident explanation it recently released that a problem from a previous network upgrade surfaced after a recent network change, leading to the outage.
The outage lasted from Nov. 12 to Nov. 13. A Microsoft service status log released by the University of Pennsylvania showed the multi-service incident (MO1550821) began on Nov. 12 and ended on Nov. 13, lasting 30 hours and 5 minutes. Microsoft managed separate logs for the Exchange Online incident (EX1550543) and the Microsoft 365 multi-service incident (MO1550821). The two logs covered service disruptions that occurred at the same time, and the cause was confirmed to be the same.
The impact was not limited to email. Exchange Online users had trouble accessing mailboxes via Outlook on the web, the Messaging Application Programming Interface (MAPI), Exchange ActiveSync (EAS) and the REST API. Mail delivery between internal corporate mail systems and Exchange Online was also affected. Some functions in Teams, SharePoint Online and Microsoft 365 Copilot that rely on Exchange Online did not work normally. Windows 365 Cloud PCs also faced access issues in environments with certain security settings applied.
Microsoft said a latent bug surfaced when some traffic was rerouted to a different path during routine network maintenance and upgrades. Exchange Online connection requests rose rapidly and piled up in queues. Processing capacity on affected servers hit its limit, leaving them unable to accept new connections. The impact spread in a chain to some services such as Teams and Copilot that depend on Exchange Online.
The company stressed that the outage was a technical problem unrelated to cyber attacks, artificial intelligence (AI) or AI agents. It said the long restoration time reflected the need to recover services in stages and verify stability at each step because demand remained very high even after the root cause was addressed.
Disruptions continued on the ground. A hospital in Toronto, Canada, told patients to contact it by phone due to the email outage, and the Calgary Board of Education was also affected across email, Teams and SharePoint services. Some police agencies in Britain asked the public to make urgent contact by phone, showing the impact also reached public services.
Microsoft apologised to customers and partners for the inconvenience and said it would release a preliminary post-incident report within 2 business days and a final report within 5 business days. It has not disclosed the exact number of users affected or details of the network upgrade through which the flaw was introduced.
The incident showed that disruptions to email infrastructure can ripple across collaboration tools, document services and AI assistant functions. Attention is on how Microsoft will bolster network change management and outage recovery procedures in follow-up reports.