As of 05:01 GMT on Friday, November 13, a global Microsoft 365 outage that began Thursday afternoon is still going, and Microsoft has not given an estimated time for recovery.
Microsoft opened Exchange incident EX1550543 at 16:17 GMT on Thursday, November 12, after users had trouble reaching Outlook mailboxes on the web, desktop and mobile. At 17:54 GMT it described the impact as being in the Americas. At 18:07 GMT, Microsoft 365 Status said the issue "may affect Microsoft 365 service connectivity for users in any region." At 18:26 GMT it opened a broader incident, MO1550821, covering the rest of the Microsoft 365 suite.
What's Broken
Exchange Online is the center of this, and that means Outlook, in the browser, on the desktop and on mobile. Microsoft also says inbound mail may be delayed.
Teams is a split picture. Microsoft says joining calls and meetings, group chats, channel messages and standard channels are largely still working. What failed are the Exchange-dependent pieces: calendars, scheduling, invites and meeting edits, Live and Virtual events, adding or syncing contacts, and creating or deleting private or shared channels.
Copilot is in the same boat. Microsoft 365 Copilot chat may not load, or may be unable to reach email, calendar and files.
Microsoft also listed a long tail of Exchange- and Graph-dependent impact, including SharePoint features that rely on Exchange (files stored in SharePoint remain available), Intune, Defender for Office 365, including Safe Links, Purview, Planner, To Do, Universal Print, Power BI and Fabric sensitivity labels, and the admin center. Power Platform advisory PP1550869 says flows, apps and Copilot Studio connectors that rely on Graph have been failing intermittently. OpenAI said new Microsoft 365 connections may be broken and tools are not working as expected. Anthropic said mail, calendar and search in its Microsoft 365 connector were failing for most requests, and that Claude for Outlook was also affected.
A separate North America Teams incident, TM1550488, ran from 15:40 to 17:09 GMT on Thursday and was fully resolved, Microsoft said. Microsoft's public Azure status page showed no Azure issues.
What Microsoft Says Went Wrong
There is no post-incident review yet, and the root cause on both main incidents is still listed as N/A.
By 19:18 GMT, Microsoft had determined that the elevated load was not coming from more traffic. A large number of servers failed automated health checks and dropped out, shifting load onto the servers that were left. At 19:06 GMT it paused all deployments and configuration changes. Later, it said a process on the affected servers that sets up secure connections "becomes overwhelmed, causing the servers to stop processing requests," and it began reverting "a recent network configuration change that correlates with the start of impact."
Recovery has not held. Servers brought back online became unresponsive again once traffic returned. At 04:10 GMT on Friday, Microsoft reported partial recovery in some regions, but said those actions were "not resulting in sustained recovery."
As of 05:01 GMT, the latest official update when this package was finalized, Microsoft said its mitigations had caused a surge of calls from routing services to backend servers, and that some machines were "stuck in a bad state," blocking a force-sync of the network change. It is scaling up virtual machines and still has no ETA.
