Microsoft 365 Outage Runs Into a Second Day as Fixes Fail to Hold

In an overnight update, Microsoft said its changes were "not taking effect as expected", because servers it had restarted became unresponsive again once they received traffic. On Friday morning it said it was "continuing to work on recovering our frontend servers", was "aggressively pursuing all pathways to reduce impact" and was optimising traffic to improve stability and capacity. It gave no time for full recovery.

The length of the outage turns a service failure into an operating problem for the many companies, banks, courts and government offices that run their daily work through Microsoft 365, which has more than 430 million users. Businesses in Britain and Europe started Friday without reliable email, and Swiss luxury group Richemont delayed the publication of its half-year results before releasing them later in the day. In the UAE, Outlook, Microsoft 365 and Teams were still at the top of Downdetector's local site on Friday morning.

The problems began on Thursday afternoon, European time. Downdetector reports for Outlook peaked on Thursday evening, with counts put at more than 130,000 by one outlet and nearly 140,000 by another. Microsoft said at first that many servers had become unavailable and failed automated health checks, which pushed their load onto the rest of its infrastructure. Later on Thursday it said it was examining "a small number of configuration updates" as a possible trigger.

Overnight the company reported an overloaded process that sets up secure connections on the affected servers and began rolling back a recent network configuration change that coincided with the start of the disruption. It has not confirmed that change as the cause. By early Friday it was dealing with a spike in traffic from its routing services to backend servers, partly caused by its own efforts to redistribute traffic, and planned to add virtual machines to absorb the load. Microsoft has said there is no indication of a cyberattack.

The disruption reaches well beyond email. According to Microsoft's service notices, as summarised by the Belgian trade publication ITdaily, Outlook on the web and in the app, Outlook.com, Teams calendars and meeting scheduling, OneDrive, OneNote, Planner, Purview, Copilot Chat and parts of the admin centre and Defender portal have been affected. Safe Links may fail to check web addresses and Intune device enrolment and compliance checks may stall. Microsoft says Teams chat, online meetings and SharePoint files have largely kept working. OpenAI and Anthropic also reported problems with services connected to Microsoft's infrastructure.

Charles Lamanna, a Microsoft executive vice president, wrote on X: "We regret that today's outages caused disruption for our customers, and we are working to resolve the problem."

Analyst Paul Spain warned that the event could become the "most severe technology outage in history" if it lasted 24 hours. Toby Murray, an academic, pointed to the "inherent dangers" of relying on a single major provider. The failure follows an Azure outage last month and a Microsoft 365 outage in North America in January. Administrators can follow the incident in the Microsoft 365 admin centre under the references EX1550543 and MO1550821.