Microsoft’s email outage, which began on Thursday and has hit governments, schools, airlines and businesses on several continents, was still being worked on Friday. The company says a new patch is stabilising servers, but it has not said when Exchange Online will be fully restored or what caused the failure.

Users first reported being unable to open Outlook around 16:00 GMT on Thursday, and within about an hour Downdetector had logged complaints from 74 countries and territories, according to Al Jazeera. Microsoft’s incident tracker lists the Exchange Online problem (EX1550543) as starting at 16:17 UTC. By Thursday afternoon US time, Outlook outage reports peaked at 130,000 on Downdetector, Forbes reported; by about 07:00 GMT on Friday, Al Jazeera counted at least 43,000 reports from 72 countries. The two numbers measure different moments, and Downdetector counts understate the true scale.

Microsoft’s own account has moved in steps. It first said infrastructure was under elevated load, then said the load was not from extra traffic but from “a large number of servers being unexpectedly unavailable,” as ABC Australia quoted it. Servers were failing automated health checks and showing high processor use, according to LazyAdmin’s reconstruction of the incident. Microsoft later said a process responsible for establishing secure connections became overwhelmed, causing affected servers to stop responding. It paused all deployments on Thursday, restarted servers and returned some to service using a different health-check method, but they became unresponsive again once traffic hit them. Early Friday it reported “positive results for server stability” after a new patch and said it was expanding the fix. At one point a change did not work as expected and the company examined other capacity options, LazyAdmin’s timeline shows. Microsoft has also begun reverting a recent network configuration change, but has not said that was the cause. A post-incident report is expected to give the definitive answer.

Microsoft says there is no sign of a cyberattack. Charles Lamanna, executive vice president for Copilot, Agents and Platform, wrote on X: “there is no indication this is the result of malicious or cyber activity.”

The impact reached well beyond inboxes. Outlook and Exchange Online were hit hardest; Teams, SharePoint, Copilot and Defender XDR were also affected, and a separate Teams incident in the Americas was resolved. Australia’s spy agency and federal police, Japan’s government IT systems, the US House of Representatives and the Philippine Supreme Court were among the organisations reporting problems, according to Al Jazeera and ABC Australia, and New York City’s emergency management department warned of delays. Connectors used by OpenAI and Anthropic that rely on Microsoft services were also affected. Mail is delayed, not lost, and users should not resend messages, LazyAdmin advised.

Analysts have offered theories, not findings. Technology analyst Rupert Goodwins told Al Jazeera that a faulty update on the administration network is a likely cause and that servers reconnecting can be swamped, failing again. Walter Pasquarelli of the University of Cambridge said servers failing health checks shifted traffic onto others, which overloaded, and saw no evidence that AI demand caused it. The outage follows a near 10-hour Microsoft 365 failure across North America in January and an Outlook lockout in late August.

For business customers, credits under Microsoft’s 99.9% monthly uptime commitment are not automatic: they must be claimed through the Microsoft 365 admin center, with the incident IDs, by December 31, according to LazyAdmin. This article is for information only.