Microsoft Outlook Down: Why Outlook, Teams and Copilot Stopped Working
A three-day disruption across Microsoft 365 triggered more than 138,000 Outlook outage reports after automated server health checks misfired.
Microsoft spent much of mid-November untangling one of its widest workplace outages of the year. Between November 12 and November 14, 2026, a rolling Microsoft 365 disruption cut off access to Outlook and Exchange Online, stalled Microsoft Teams chats, and knocked Microsoft 365 Copilot, SharePoint, Planner, To Do, and Power BI offline for organizations around the globe. At the peak of the incident, Downdetector logged more than 138,000 user reports for Outlook alone.
Because the breakdown hit corporate email and collaboration tools simultaneously across North America, Europe, and parts of Asia, IT administrators initially worried about a targeted intrusion. Microsoft quickly ruled out a cyberattack, pointing instead to an internal infrastructure cascade inside its own server fleet.
How a health-check glitch snowballed across Microsoft 365
According to Microsoft's service health disclosures, the root cause began with server-side infrastructure health-check failures. In hyperscale cloud environments, automated watchdogs continuously probe servers and pull unresponsive machines out of rotation so live user traffic routes elsewhere. During the November outage, faulty health-check behavior caused healthy or lightly stressed infrastructure to be flagged and removed from service.
Once a large share of servers dropped offline, the surviving pool absorbed a sudden surge of authentication requests, mailbox syncs, and Teams presence pings. That overload slowed response times further, tripping additional health checks and creating a self-reinforcing loop. Because Microsoft 365 Copilot, Planner, and SharePoint rely on shared Exchange Online identity and data layers, the collateral damage spread far beyond email.
Why recovery took days instead of minutes
Bringing an overloaded global cloud service back online is rarely as simple as flipping a switch. Engineers had to isolate the faulty health-check logic, restore drained server capacity in stages, and throttle incoming traffic so millions of retrying desktop and mobile clients did not immediately crash the newly recovered nodes. While core mail flow returned first, lingering delays in Teams notifications and Copilot queries persisted into November 14 as queued background jobs cleared.
For enterprise IT teams, the episode highlights a familiar operational risk. Even when defenses hold, tightly linked SaaS suites share failure domains, meaning one automated bug can stall email, chat, documents, and analytics together.