Global Microsoft 365 outage drags into a second day

A widespread outage at Microsoft knocked out Outlook email, Microsoft Teams and several other cloud services for customers around the world beginning Thursday afternoon, and the company had not achieved a full recovery by early Friday, more than 10 hours after the disruption began.

Microsoft's own status updates, relayed by multiple news organizations, said the outage started shortly after 1 p.m. ET on November 12 and was tracked internally under an admin-center advisory numbered MO1550821, covering Exchange Online, Microsoft Teams, SharePoint Online, Microsoft 365 Copilot and Microsoft Defender XDR, according to BleepingComputer's reporting. NBC News counted more than a dozen affected cloud services in total. By the time NBC and ABC News updated their accounts just before and just after midnight, Microsoft was still describing the problem as unresolved, and both outlets flagged their stories as developing.

What went wrong

Microsoft's explanation shifted over the course of the day. It first attributed the disruption to elevated infrastructure load across multiple regions, then said its engineers had determined that servers were becoming "unexpectedly unavailable," straining the infrastructure that remained online, according to BleepingComputer's account of the company's updates. Later, Microsoft narrowed the diagnosis further, saying "a core process on affected servers" was becoming overwhelmed, which caused a broader loss of server capacity, per NBC News' reporting of the company's statements. Engineers paused all software deployments, redistributed network traffic worldwide and began manually restarting affected servers. Microsoft also said it was testing whether reverting a small number of recent configuration changes would speed the fix, without specifying what those changes involved or when they had been made.

The remediation did not hold. By around 7 p.m. ET, Microsoft said it had returned a portion of affected servers to service and that they "remain stable so far," BleepingComputer reported. But by shortly before midnight, the company reported only "partial recovery across regions," said the fixes were not producing "sustained recovery," and warned that servers it had already restored "can become unresponsive again once they receive traffic," according to both NBC's and ABC's accounts of the company's status messages. Azure, Microsoft's separate cloud-computing platform, was not affected, NBC reported, which meant the outage was concentrated in the company's productivity and communications software rather than in the infrastructure layer it rents out to other businesses and competing cloud customers.

A global, multi-service disruption

The outage's reach went well beyond email. Reported impacts included Teams messaging and calling, SharePoint, Microsoft Planner, Microsoft To Do, Windows 365 virtual desktops, cloud printing and mobile device management on the enterprise side, as well as consumer services including Outlook.com, OneDrive, OneNote and the web version of Microsoft 365, according to NBC News. A separate, shorter Teams incident earlier Thursday morning, tracked by Microsoft as TM1550488 and reported fully resolved by 10:53 a.m. ET, had already caused message delays and errors creating or retrieving chats; the company has not said publicly whether the two incidents share a common cause, BleepingComputer reported.

Governments and public services felt the disruption alongside private businesses. New York City's Office of Emergency Management warned residents that city email and other online services relying on Microsoft systems could see delays, ABC News reported. NBC News reported that federal agencies, offices on Capitol Hill, and government bodies in California were among those affected, and that a Colorado sheriff's office had to process bond payments in person because its systems were down. A delivery-truck manufacturer, Workhorse Group, put off its scheduled earnings report as a result, according to NBC News. None of these downstream effects were disputed by Microsoft in the material reviewed for this article, though the company did not respond directly to NBC's request for comment.

"We appreciate our customers' patience while we work to restore services as soon as possible," a Microsoft spokesperson told Bloomberg, as relayed by NBC News.

A recurring problem for the world's back office

Microsoft 365 underpins email and document collaboration for a very large share of the world's offices, schools and government agencies, and NBC News reported that the disruption halted mail flow for many of the roughly 450 million users the company has said rely on its email and productivity suite. That scale is also why outages of this kind draw outsized attention: unlike a single consumer app going dark, a Microsoft 365 failure can simultaneously stall internal communication at banks, hospitals, police departments and city halls that have no ready substitute for their primary email and chat system, and no easy way to notify staff of the problem when the notification system itself runs on the same affected software.

Thursday's episode was not an isolated event. A separate Microsoft 365 outage in January dragged on for almost ten hours before the company confirmed a fix, after it traced the problem to "a portion of service infrastructure in North America that is not processing traffic as expected," The Register reported at the time. Exchange Online has also had shorter, more contained incidents since, including episodes in August and twice in early November. Each has had a distinct technical explanation, but the recurrence has fed a broader debate among corporate IT departments about how much of their operations should depend on a single cloud provider's infrastructure, and what backup communication plans they need for when that provider goes down. That debate predates this week's outage, but incidents of this scale and duration tend to sharpen it each time they occur.

What happens now

Administrators with access to the Microsoft 365 admin center can track an active incident's status through defined stages, from "investigating" through "restoring service" to "service restored," and Microsoft commits to eventually publishing a post-incident report that lays out a root cause and corrective steps, according to Microsoft's own documentation on the service health dashboard. Those detailed reports, however, are generally visible only to administrators rather than the public, and customers without admin access are directed instead to Microsoft's public status page or its Microsoft 365 Status account on the social-media platform X for updates during an active incident.

As of the most recent public updates early Friday, Microsoft had not given a timeline for full restoration and said it was continuing to evaluate additional steps to stabilize the affected infrastructure, according to ABC News. Customers who rely on Exchange Online and Teams for business-critical communication are likely to press the company for a fuller account of why servers became unexpectedly unavailable in the first place, and why its initial remediation repeatedly failed to hold through the day. A published post-incident report, if and when one follows, would be the first place that fuller explanation is likely to appear, and it would add to a lengthening record of similar reports Microsoft has issued after previous Microsoft 365 disruptions this year.