Why did Microsoft go down worldwide? Outage causes explained - The Jerusalem Post

While the company hasn't issued a final resolution, and the problem reached Asia-Pacific on Friday, most experts ruled out the possibility that the incident was the result of a cyberattack. According to official Microsoft statements, the company was investigating a possible internal error that prompted accounts to be disconnected from North American servers, causing a domino-like effect when those accounts were moved to other servers globally.

Server failure investigation

In technical terms, the first assessment indicates that some of Microsoft's servers failed a health check, prompting the connection disruption and redistribution of users to servers in other parts of the world. Experts suggest that because the issue wasn't a cyberattack, Microsoft will likely need to identify the exact server that failed the health check and triggered the massive outage.

Why not a cyberattack

Three facts push the cyberattack theory aside. Firstly, Microsoft's assessment describes solutions as an internal investigation into its own server issues rather than an external attack. Second, no spike in users indicated a Distributed Denial-of-Service (DDoS) attack, with most reports showing normal user levels. Third, no hacker group claimed responsibility for the attack even after services were down for several hours.

Global impact

According to Downdetector, the Microsoft incident reached servers in Israel, with disruptions detected in Tel Aviv, Jerusalem, and Haifa. The outage affected Outlook, while Microsoft 365 services and Copilot, Microsoft's AI assistant, also faced similar disruptions. The incident has been active for over 12 hours worldwide, with users in the Asia-Pacific region reporting disruptions on Friday.

Initial estimates suggest the surge in traffic to alternative servers caused delays and further outages, with some companies having to send employees home early due to persistent connection issues.

Recent precedents

In January, a similar but smaller outage lasted almost nine hours, with about 3,000 reports on Downdetector. That disruption was triggered when one of Microsoft's data centers in Cheyenne, Wyoming, was taken down for maintenance, causing a cascading effect similar to the most recent one.

Last October, a massive disruption of Amazon Web Services (AWS) knocked out thousands of websites along with popular apps like Snapchat and Reddit. That incident was also described as an internal error rather than a cyberattack.