Microsoft's Global Outage Stretched Into a Second Day. Here's What Broke — and What Didn't

Millions of people opened their inboxes on Thursday and found nothing. No new mail, no calendar invites, no Teams meetings loading. By Friday morning, the problem was still there.

Microsoft's global outage began on Thursday, November 12, when connections to Exchange Online — the cloud email service behind Outlook — started failing at around 16:17 UTC. Within hours it had spread across Microsoft 365: Outlook, Teams, Word, Excel, PowerPoint, SharePoint, OneDrive and Copilot all took hits. Downdetector, which tracks outages through user reports, logged nearly 130,000 complaints about Outlook alone at the peak, and reports came in from 74 countries.

It was not a small glitch, and it was not a quick one. The disruption ran for more than five hours on Thursday and carried into Friday, November 13, with thousands of users still reporting problems as the workday began in the United States. Some Teams chats and meetings came back for many users around 02:00 UTC on Friday. Others were still waiting.

Microsoft has been blunt about one thing: this was not a cyberattack. Charles Lamanna, the company's executive vice president for Copilot, Agents and Platform, apologized to customers and said there was no indication the outage came from malicious activity. A Microsoft spokesperson told The Guardian the same. Australian ministers weighed in with the same message.

So what actually broke? Microsoft traced the trouble to its own servers. A process responsible for establishing secure connections became overloaded, and affected servers started failing automated health checks. When a server fails a health check, it drops out of service. Enough of them dropped out at once that the remaining machines buckled under the shifted load, and the cascade spread across regions.

The company first pointed to unusually high load on parts of its infrastructure. It later corrected itself: the load was not driven by a surge in users. The problem was that a large number of servers had unexpectedly become unavailable, straining what was left. In Microsoft's own words, "we determined the elevated load isn't driven by an increase in traffic, but by a number of servers being unexpectedly unavailable, straining the remaining infrastructure."

Fixing it has proved harder than expected. Microsoft said it paused all deployments, moved traffic off the affected infrastructure and began restarting servers to bring them back online. Some of those servers stopped responding again once traffic returned, which is why recovery has dragged on. The company is testing a configuration change aimed at the overloaded process, and it has published no firm timeline for full restoration. It is tracking the Exchange Online incident as EX1550543 and the broader Microsoft 365 disruption as MO1550821.

The outage reached beyond Microsoft's own apps. OpenAI and Anthropic both reported problems with their Microsoft 365 connectors during the disruption, a reminder of how much of the modern workplace runs through a single provider's plumbing. When one company's email and calendar go down, meetings stall, documents sit unopened and whole offices effectively pause.

That scale is the real story here. Microsoft 365 is the default productivity suite for governments, hospitals, schools and businesses around the world. A failure in its core does not stay a tech story for long. It becomes a story about missed deadlines, delayed decisions and a Friday morning spent staring at an error message.

Microsoft has not given a definitive explanation for why the disruption lasted so long. For now, the official line is infrastructure and configuration problems, not an attack — and a recovery that is still under way.