A Microsoft 365 outage lasting more than 22 hours left millions of people in over 70 countries without email and work tools. According to the BBC, it began with routine automated checks that took some servers offline, overloading the remaining ones until they failed too. Recovery was gradual and took almost a full day.
The outage hit Outlook, Teams, SharePoint, Copilot and Defender. Microsoft 365 has around 430 million users worldwide. The fallout was felt in very different places: the US House of Representatives lost email access, Australian intelligence agencies had the same problem, courts in Singapore were slowed and some airline support desks in India were blocked. In the US, several city administrations asked citizens to call offices by phone instead of writing. In Canada, a network of Toronto hospitals declared a "code grey", an infrastructure emergency, though it said patient care was not affected.
Not a cyber-attack
Microsoft apologised and ruled out a cyber-attack, describing a "technical issue" linked to a recent network upgrade, according to Windows Central. The Guardian reports the company also said the failure was not related to AI agents. In the following hours Microsoft reported improvement, with email queues "significantly reduced and approaching normal levels", while warning some users could keep seeing intermittent problems.
This is not the first time a single provider has taken essential services down with it. In August an outage at GitHub, also owned by Microsoft, had blocked software development worldwide for hours. The question of depending on a few large providers has also returned to the European debate, with some lawmakers outside the United States arguing episodes like this show the risks of overreliance on American tech.
How much the dependence weighs
The open question is how prepared organisations are to lose their tools for a full day. For small and medium businesses, as Flora Hamilton of the Small Business Charter points out, email is the basis of every relationship with customers and suppliers. Gergely Orosz, a former senior engineer at several large tech companies, notes that outages are not unusual, but that the time taken to return to normal was unusually long. One possible fix, he suggests, would be splitting services by region, as Amazon already does with AWS, so a problem stays contained.
For now Microsoft is reviewing what went wrong, and the findings should help prevent a similar episode.

No comments yet. Be the first!