Microsoft experienced one of its biggest global outages on Thursday, with thousands of clients worldwide reporting they could not connect to services such as Outlook, Microsoft 365 and Teams. The disruption affected companies and some government operations globally, and more than 130,000 users reported issues on Thursday, though reports have since decreased while many users still report problems.
Microsoft has not issued a final resolution, and the problem reached the Asia-Pacific region on Friday. According to official Microsoft statements, the company was investigating a possible internal error that caused accounts to be disconnected from North American servers, followed by a domino-like effect as those accounts were moved to other servers around the world. The surge in traffic to those servers caused delays and further outages, and some companies had to send employees home early because of the persistent connection problem.S1
Technically, the first assessment indicates that some servers failed a health check, prompting the connection disruption and the redistribution of users to servers in other parts of the world. Experts say Microsoft will likely need to identify the exact server that failed the health check and triggered the outage. The incident also reached Israel, with disruptions reported in Tel Aviv, Jerusalem and Haifa, affecting Outlook as well as Microsoft 365 services and Copilot, Microsoft's AI assistant. The incident had been active for over 12 hours worldwide by the latest reports.S1
Most experts have ruled out a cyberattack. Microsoft's assessment describes its response as an internal investigation into issues with its own servers rather than an attack from outside, no spike in users pointed to a Distributed Denial-of-Service (DDoS) attack, and no hacker group claimed responsibility, even after service was down for several hours. A full explanation will depend on a public status report and the completion of Microsoft's internal investigation.S1
The incident follows a similar, smaller outage in January that lasted almost nine hours and drew about 3,000 reports on Downdetector. That disruption was attributed to one of the data centres in Cheyenne, Wyoming, being taken down for maintenance, with a cascading effect similar to the current one. Last October, a massive disruption of Amazon Web Services knocked out thousands of websites.S1
