Microsoft’s outage, which took out Outlook, Teams and Copilot on Thursday, ran into Friday, with the company saying its early recovery steps had not delivered the expected relief. In its latest posts, Microsoft said it was seeing early stabilisation and capacity improvements as recovery actions continue, but it had not declared the incident resolved or given a time for full restoration. Reports from across the globe suggest disruption across as many as 74 countries, though the number could be higher. Microsoft says it will publish the root cause and “defense-in-depth repair items” in a post-incident report once recovery is complete.
The Exchange Online incident, tracked as EX1550543, began around 11:17 a.m. ET on Thursday. The wider incident, MO1550821, opened about two hours later and listed Exchange Online, Teams, SharePoint Online, Copilot 365 and Defender XDR, with Intune, Defender for Office 365 and some Purview functions added later. Downdetector logged nearly 130,000 Outlook reports at the first peak.
Microsoft’s account of the fault has narrowed since Thursday. It first cited elevated load, then said the load was not driven by more traffic but by a number of servers becoming unexpectedly unavailable and straining the rest, with the affected machines failing automated health checks. An alternative health-check method returned some servers to service, and around 6 p.m. ET Microsoft said they were stable. It later said they could become unresponsive again once they received traffic, and that it was analysing diagnostic data to find the trigger. Microsoft then said a core process on the affected servers was becoming overwhelmed and causing capacity loss, and Forbes reported that the process handles establishing secure connections.
The recovery work has included restarting servers in isolated groups, temporarily limiting which types of traffic can reach the affected infrastructure, scaling up virtual machines, and applying a patch that Microsoft says is improving stability. Al Jazeera reported that Microsoft was also rolling back a network change on Friday, without saying whether it was the cause. Microsoft said its latest traffic-optimisation actions were producing positive results and were being applied across the affected regions as it works toward persistent recovery, and CNN reported its own phrase for the repairs: “marginal success.”
To preserve mail delayed by the outage, Microsoft extended its email timeout to 48 hours. Charles Lamanna, Microsoft’s executive vice president for Copilot, Agents and Platform, apologised for the disruption and said there was “no indication this is the result of malicious or cyber activity.”
The outage also reached AI tools that read Microsoft 365 data. Anthropic said it had identified an issue with an upstream provider affecting its Microsoft 365 connector, and OpenAI’s connector also reported problems, according to DW.
It follows an Azure and Microsoft 365 outage on October 29 last year and a January 22 incident that Microsoft attributed to part of its North American service infrastructure not processing traffic as expected. Administrators can follow the current incident in the Microsoft 365 admin center under EX1550543 and MO1550821, or at status.cloud.microsoft when the admin center is unavailable.