The company said measures intended to stabilise its infrastructure had failed to deliver the expected improvement, forcing engineers to explore alternative ways of restoring capacity. The disruption continued into Friday in parts of the world, with no confirmed timetable for complete recovery.
“We’re identifying alternate workstreams that can be applied to increase capacity in the face of sustained traffic management imbalances,” Microsoft said in a service health update.
The problems centred on Exchange Online, the cloud platform supporting email, calendars and related functions. Microsoft said services dependent on Exchange Online were affected, although features operating independently of that infrastructure generally remained available.
Outlook users encountered difficulties opening mailboxes, sending and receiving messages and connecting through different applications. Microsoft Teams, SharePoint Online, Microsoft 365 Copilot and Microsoft Defender XDR were also among services identified as potentially affected.
Downdetector recorded more than 130,000 reports of Outlook problems in the United States around 1pm Eastern Time on Thursday. More than 25,000 reports were still appearing at approximately 9.30pm, indicating that substantial disruption persisted despite attempts to restore service.
Reports of difficulties also emerged across the Asia-Pacific region, including Australia, Japan and India, as the disruption extended beyond the American working day. Downdetector figures reflect user-submitted complaints and do not establish the number of individuals or organisations affected.
Microsoft initially identified elevated loads across infrastructure handling customer connections. Its investigation subsequently found that numerous servers had become unexpectedly unavailable, placing additional pressure on those still operating. The company said the elevated load was not caused by an overall increase in customer traffic.
Engineers redistributed connections across the company’s worldwide infrastructure, diverted traffic away from affected equipment and restarted servers. Some machines returned to service, but Microsoft later observed that they could become unresponsive again after receiving incoming requests.
A further investigation identified a process responsible for establishing secure connections that was becoming overwhelmed and preventing affected servers from processing requests. Microsoft began validating a configuration adjustment designed to address that behaviour while also examining whether a network configuration change contributed to the disruption.
The company said it was reversing that network change and restricting certain types of traffic in selected locations to protect vulnerable infrastructure. It also tested rebuilding some affected servers using an earlier software configuration. These measures were intended to restore service progressively without placing additional strain on functioning systems.
However, a subsequent service update acknowledged that the changes intended to improve reliability were not producing the expected results. Microsoft said it was examining traffic flows and server behaviour to identify further adjustments capable of improving stability.
The outage affected organisations dependent on Microsoft 365 for everyday communication. Australian government agencies and the United States House of Representatives were among institutions experiencing problems, while disruptions to Microsoft 365 connectors were also acknowledged by artificial intelligence companies OpenAI and Anthropic.
Workhorse Group postponed the announcement of its quarterly financial results, citing the disruption to its business operations. The decision illustrated how failures affecting widely used workplace software could interrupt activities beyond email correspondence and online meetings.
Microsoft’s service advisory distinguished between affected and unaffected Teams functions. Core capabilities including joining meetings, group conversations and channel messages largely remained operational, but calendar scheduling, invitations and meeting changes could fail because of their dependence on Exchange Online.
Other potentially affected functions included certain Copilot responses, Microsoft To Do operations, SharePoint features linked to email and some administrative or compliance services. Microsoft said files stored directly in SharePoint remained accessible, underscoring that the interruption did not disable every component of its cloud software suite.
A separate Teams messaging incident earlier on Thursday affected users whose connections passed through infrastructure in North and South America. Microsoft said that issue, attributed to resource contention, had been resolved. It did not establish a connection between that incident and the broader Exchange Online disruption.
The company also paused deployments and configuration changes while reviewing possible contributing factors. Its service notices said users might see gradual improvements as restored servers accepted connections and applications re-established access to their mailboxes. The company had not provided a definitive count of affected customers or a firm restoration deadline by its update.
Follow Arabian Post
Select Arabian Post as your preferred source on Google and MSN News for trusted business news and Arab politics and updates.
