Microsoft is working to stabilize its 365 services after an outage that began in the United States on the evening of November 12 and has since rolled across time zones, leaving workers in many countries unable to reach their email, calendars, and, in some cases, their AI assistants.
Outlook and Exchange Online took the first hit, with Microsoft initially pointing to trouble reaching mailboxes in the Americas before warning that connectivity could suffer in any region.
Downdetector received reports from 74 countries and territories within about an hour, and nearly 130,000 US Outlook complaints were logged shortly after the incident began, although those figures count reports rather than confirmed individual users.
Microsoft’s explanation is that a process responsible for setting up secure connections on affected servers became overwhelmed and stopped handling requests.
Engineers have reversed a related network configuration change to check whether it played a part, tested fixes on a subset of servers, restarted infrastructure, and begun feeding capacity back in gradually, which explains why some users are back online while others still see delays.
The damage reaches well beyond email, with Teams functions that lean on Exchange, such as calendars and meeting scheduling, faltering even where core chat and calls keep running, while Copilot has struggled to reach work emails and files. SharePoint, Planner, Intune, Power BI, and Windows 365 are also on the affected list, and both OpenAI and Anthropic reported connector problems tied to Microsoft 365.
Businesses felt it quickly, with Workhorse Group postponing its quarterly results announcement and citing disruption from the outage, while reports pointed to trouble at courts, schools, and government bodies.
Microsoft has not blamed an attack, and one of its executives said there was no sign of malicious activity, though that reflects the investigation so far rather than a final verdict. Security teams are nonetheless warning of phishing aimed at confused users, with fake restoration links the likely bait.
READ: Microsoft Shuts Down Africa Development Centre in Nairobi
For Kenyan companies and agencies that run their daily work through Microsoft 365, the episode makes a case for keeping a backup communication channel outside the same tenant, along with local copies of key contacts. It also pays to resist hammering sign-in pages while Microsoft recovers, since repeated retries can add to the load on already strained servers.
Microsoft’s status page still shows degradation, so full recovery looks set to be gradual, and its post-incident review will show how a single connection fault managed to reach this far.
