For millions of office workers, the working week ended with a spinning wheel instead of an inbox. A sprawling Microsoft failure that started on Thursday afternoon, US time, knocked out Outlook and hobbled Teams, SharePoint, OneDrive, Copilot and other parts of the Microsoft 365 bundle, and it was still not fully fixed by Friday morning in Europe.
Downdetector logged close to 140,000 complaints about Outlook at the peak, and the tracking site said people in more than 70 countries were affected, from South Africa and India to New Zealand. Because the trouble began around lunchtime on the US East Coast, it rolled straight into the Asian business day and then into breakfast time across Europe. Japanese companies and public bodies were among those hit.
Microsoft has offered a changing account of what went wrong. Early updates blamed heavy load in several regions. The company then stressed that traffic had not actually risen; instead, a batch of servers had dropped out unexpectedly, leaving the remaining machines to carry far more than they were built for. Engineers later traced the failures to the automated system that decides whether a server is healthy enough to take work, saying a core process on the affected machines had become “overwhelmed.”
The company says there is no sign of a cyberattack. Independent specialists agree that a self-inflicted fault fits the facts better: when the watchdog itself misjudges, it pulls good servers offline and the rest buckle.
To dig out, Microsoft rebooted hardware, froze software rollouts and pushed connections onto other data centers. By early Friday it reported “early signs of stabilization” and most Teams access had returned, but it gave no timetable for a complete recovery of email. Some 430 million people rely on its 365 apps.
