The Microsoft 365 outage that began on Thursday evening had still not been resolved by Friday morning. Outlook, Teams, SharePoint and Copilot are suffering disruptions worldwide, and servers that Microsoft restores are failing again.
The problems began on Thursday at 5.17 p.m. Belgian time at Exchange Online, the email service behind Outlook for businesses. Around 7 p.m., Downdetector counted more than 130,000 reports about Outlook, and within the hour the platform received reports from 74 countries and territories. Emails are also not arriving at the ITdaily editorial office this morning.
Administrators can follow the incident in the Microsoft 365 admin center under EX1550543 and MO1550821. If the admin center does not load, the updates are available on Microsoft’s public status page.
Microsoft 365 outage also hits Intune
Users may be unable to reach their mailbox via Outlook on the web, the Outlook app, ActiveSync or the REST API, and email may arrive with a delay. Other services falter as soon as they need Exchange Online. In Teams, the calendar and the scheduling of meetings may fail. According to Microsoft, chat, online meetings and files in SharePoint Online largely work.
Copilot Chat may not load and cannot access email, calendar and files. Intune, Purview, Planner, To Do and parts of the admin center and the Defender portal are also affected. Safe Links in Defender for Office 365 may be unable to check URLs when someone clicks a link in Outlook or Teams. Anthropic and OpenAI reported that the Microsoft 365 connectors for Claude and ChatGPT are not working.
Restored servers fail again
“The increased load is not caused by a rise in traffic, but by a large number of servers being unexpectedly unavailable, which shifts the load to the rest of the infrastructure”, Microsoft wrote on Thursday evening. The servers dropped out of service because they failed automated health checks.
Around 9 p.m., Microsoft had not yet found a recent change that coincided with the start of the outage. An hour later, it was investigating “a small number of configuration updates” as a possible cause. During the night, the company reported that a process that sets up secure connections on the affected servers is becoming overloaded. It then began rolling back a recent change to the network configuration “that coincides with the start of the impact”.
Servers that Microsoft restarted became unresponsive again as soon as they received traffic. On Friday around 4 a.m., the company wrote that the changes it had made were “not taking effect as expected”. According to Microsoft, there is “no indication” that malicious activity is behind the outage.
On Friday around 6 a.m., Microsoft reports a spike in traffic from its routing services to the backend servers, partly caused by its own redistribution of traffic. Additional virtual machines should reduce the load. “We are running into problems because the machines are stuck in a bad state, which means our changes are not synchronizing as expected”, the company writes.
In January, Microsoft 365 already suffered an outage in North America lasting almost ten hours, and at the end of August, thousands of Outlook users worldwide could not access their mailbox.
Microsoft has not yet given a time for full recovery. The company still has to confirm whether the network change it is now rolling back caused the outage.