Microsoft said it was restoring some affected servers late Thursday afternoon after widespread service disruptions prevented users from accessing Outlook and other applications. The company reported limited recovery in some regions as it redirected traffic and worked to bring unavailable infrastructure back into operation.
Nearly 130,000 users reported Outlook problems on Downdetector shortly after 1 p.m. ET. Reports declined after that peak, but users continued describing difficulties sending and receiving messages or loading their inboxes. Thousands also reported problems involving Teams, Copilot and other Microsoft 365 services.
Microsoft said the disruption affected access to Exchange Online mailboxes through multiple connection methods in the Americas. Its investigation identified elevated load on part of its infrastructure, alongside servers that had become unexpectedly unavailable after failing automated checks. The company said the increased load was likely contributing to connection errors.
To reduce pressure on the affected systems, Microsoft introduced protective traffic management measures at a specific location. It said those measures could be expanded if they proved effective. The company also shifted traffic and connections to alternative routes where possible.
Recovery was expected to be gradual as capacity returned and users’ applications reconnected. Earlier in the day, Microsoft said it had resolved a Teams issue in North America, although some users continued reporting connection problems.
In Canada, Via Rail warned that some passengers might not receive emails about their journeys because of the outage. The rail operator directed customers to its trip tracker or SMS notifications for updates. CBC News said it had requested additional information from Microsoft about the disruption’s impact on Canadian users.
By late Thursday afternoon, Microsoft had described partial restoration rather than a complete return to normal service. Its updates indicated that recovery varied by region, while reports of email and application access problems had continued after the peak.