Microsoft said connection problems on Thursday affected Teams, Outlook and its AI software Copilot. Some servers unexpectedly became unavailable, increasing the load on the remaining infrastructure. The company said users in any region could be affected and that it was testing a configuration change to restore service. The disruption may have briefly affected businesses and individuals relying on the tools for communication and daily work; Microsoft has not disclosed how many users were affected or how long the problems lasted.
Reports of connection problems in the US and UK
Microsoft initially reported issues in the US, then said the connection problems could affect users in any region. On Thursday evening, reports of problems with Microsoft products rose sharply among UK users on the outage-tracking site Downdetector.
In the US, Downdetector logged more than 131,000 reports of Outlook problems at around 18:15 GMT. Those reports were submitted by users and do not represent a Microsoft-confirmed count of affected accounts. Reports in both the US and UK indicated that the issue was not confined to one market.
Microsoft links disruption to unavailable servers
Microsoft said parts of its infrastructure were experiencing higher loads, but not because of increased traffic. Several servers had unexpectedly become unavailable, putting more pressure on the remaining infrastructure and triggering connection errors.
In a later update, the company said the affected servers had failed automated health checks and were therefore marked as unavailable. Microsoft said all affected servers had been restarted. At 19:50 GMT, it said it was testing a configuration change to address the problem.
Full recovery and scope remain unclear
Microsoft has outlined the suspected cause and the remediation under way, but has not said when all services will return to normal or confirmed the full scale of the disruption. Simultaneous connection problems with Teams, Outlook and Copilot could interrupt business communications, email and access to AI tools.
The company has attributed the incident to unexpected server failures and automated health checks, and said the affected servers have been restarted. The extent of recovery and how user-submitted reports compare with the number of accounts actually affected remain unconfirmed.