Microsoft was working Friday to restore access to Outlook, Teams and other Microsoft 365 services after a disruption that spread across regions and affected businesses and government agencies. The company reported signs of stabilization as it worked to restore server capacity, the Associated Press reported. The Journal reported partial recovery in areas where Microsoft had made adjustments, leaving the extent of restoration across customers uncertain.

The incident began Thursday and continued for more than 12 hours, according to Reuters. Microsoft’s public status page posted information about the disruption because some administrators could have trouble accessing the regular Microsoft 365 admin center. That meant the incident affected both the services people use and one of the channels administrators normally use to find out which parts of those services are down.

Exchange failures reached beyond email

A Microsoft incident notice reproduced by the Penn said features that depended on Exchange Online and email services were affected. Outlook users could have difficulty reaching their mailboxes through web and desktop applications, and email delivery could be delayed. The notice also listed problems with the calendar and meeting scheduling in Teams, where those features rely on Exchange. Features independent of Exchange continued to operate, according to the notice.

That dependency explains how one incident could appear differently across a workplace. An employee might be able to open an application and still be unable to retrieve a calendar or schedule a meeting. The same advisory said some Microsoft Graph requests that needed mail, calendars or contacts could time out. An application that uses those requests can consequently show a failure even when its own interface is available.

Restoring capacity took several steps

Microsoft told customers it was returning affected servers to service and moving some traffic to alternate routes. The notice reproduced by Penn said the company was also applying traffic controls in a specific location to reduce pressure on the affected infrastructure. Microsoft said it would expand those measures if they showed positive results. As capacity returned, customers could see gradual improvement while their applications reconnected.

Microsoft told the AP that some earlier changes had fallen short of expectations. Reuters reported that the company later saw signs of stabilization and said it had found no evidence of malicious activity. The status updates described how the company was trying to bring capacity back. The available notices did not establish a complete explanation of why the incident started or show when each affected customer could expect to regain normal service.

Dependent services reported their own disruptions

Anthropic said the Microsoft incident caused failures in mail, calendar and search tools used by its Microsoft 365 connector. The AP and The Wall Street Journal also reported that OpenAI was investigating problems with Microsoft connectors. These tools use Microsoft services to carry out tasks involving work data. When a dependent mail or calendar service is slow or unreachable, a connector may fail even though the assistant can still answer other requests.

Outside the software companies, the interruption complicated routine business. Reuters reported disruptions to some financial firms in Asia and problems sending earnings communications in Europe. One Taiwanese company cancelled a bond repurchase and another confirmed trades by fax, Reuters reported. Those examples describe specific responses by affected firms. They do not establish a count of companies affected or a total cost for the outage.

Microsoft’s notice also described effects inside its own suite beyond Outlook and Teams. Some users could have trouble viewing their recently used files, while features in Microsoft 365 Education such as Assignments could also be affected. The notice listed potential problems with a tool that verifies links in Outlook and Teams. These examples illustrate why administrators following the incident notice had to look at individual features as well as the headline availability of each application.

Recovery can differ by customer and feature

The incident’s scope makes an early improvement report hard to translate into a single all-clear. Microsoft’s notice said elements of the same application could have different dependencies. A customer whose email starts moving again may still need to check whether calendar updates arrive, meetings can be scheduled or third-party tools can access work data. Administrators also have to distinguish delayed messages from messages that need to be resent, since the notice described delivery delays alongside access failures.

The next test is whether those individual functions hold up as Microsoft returns servers to normal traffic. The company’s public dashboard and its more detailed incident notices provide separate views of recovery. One shows whether a wider disruption is active and the other describes the functions that may still be impaired. The amount of disruption and its duration can differ across customers, so a company-wide statement of improvement is a starting point for assessing recovery. Full restoration is best established through successful use of the functions that failed and a closing update from Microsoft.