Hyderabad: A Microsoft 365 outage on Thursday disrupted email, messaging, file storage and AI tools for hundreds of thousands of users across several regions. Microsoft traced the disruption to internal server failures rather than a sudden increase in user traffic.
The disruption began in the Americas before spreading to other regions. Meanwhile, outage-tracking website Downdetector recorded nearly 130,000 reports for Outlook shortly after 1 p.m. Eastern Time. Microsoft 365 received tens of thousands of additional reports.
Teams, Copilot, OneDrive and SharePoint also saw sharp increases in complaints. At one stage, Downdetector received nearly 20,000 reports per minute from dozens of countries.
Reports began declining after Microsoft started deploying fixes. However, many users continued facing problems sending or receiving emails and accessing their inboxes throughout the afternoon and evening.
Microsoft tracked the Exchange Online disruption under incident ID EX1550543. It assigned MO1550821 to the broader Microsoft 365 service disruption.
The company’s advisories identified Exchange Online, Microsoft Teams, SharePoint Online, Microsoft 365 Copilot and Microsoft Defender XDR as potentially affected services. Microsoft also warned that users in any region could encounter connection problems.
Updates through the Microsoft 365 Status account and admin center outlined the investigation. Initially, engineers detected unusually high loads across parts of the company’s infrastructure. However, further checks confirmed that user traffic had not increased.
Instead, numerous servers had unexpectedly stopped operating after failing automated health checks. This reduced the available capacity across the affected infrastructure.
As a result, the remaining servers had to manage normal traffic without sufficient support. The additional strain caused connectivity errors and disrupted access to Microsoft services.
Microsoft subsequently identified a specific component involved in processing server health information as a possible source of the failures. Engineers tested an alternative health-check method that successfully restored some affected servers. The company then began extending the change across more systems.
Meanwhile, Microsoft paused deployments and redistributed traffic across sections of its global network. Engineers also began restarting servers that had stopped responding.
Later investigations identified an overloaded core process on some servers, further contributing to the loss of capacity. However, Microsoft said it had found no recent change associated with the beginning of the incident.
AWS confirms normal operations as Microsoft expands recovery
Microsoft services dominated Downdetector’s list of the most-reported disruptions during the incident. Outlook, Microsoft 365, Copilot, Teams and OneDrive recorded substantial activity.
Other Microsoft products also appeared on the list, including Microsoft Store, Xbox Network and Microsoft Account. Meanwhile, smaller spikes emerged for Gmail, Google, Amazon Web Services, Facebook, Zoom and Cloudflare.
Such secondary reports can occur during major technology outages. Users often check unrelated platforms when widely used services stop responding. In addition, tools connected to affected platforms can experience problems of their own.
However, Amazon Web Services said its systems continued operating normally throughout the disruption. The company indicated that some reports appeared connected to Microsoft’s separate service problems.
Google Cloud status pages also showed no comparable service failures. Therefore, the additional reports did not establish a broader outage across major cloud providers.
Smaller numbers of complaints appeared for Blizzard Battle.net and Overwatch. However, those reports remained limited and had no identified connection to the Microsoft incident.
By late Thursday, Microsoft confirmed that some servers affected by the Microsoft 365 outage had returned to operation. The company continued expanding its recovery measures across the affected infrastructure.
Nevertheless, full service had not been restored, and Microsoft provided no definite timeline for complete recovery. The company directed administrators to the Microsoft 365 admin center and status.cloud.microsoft for further updates.