A World on Standby
On Thursday, November 12, 2026, work slowed to a crawl for millions around the globe as Microsoft 365 services began to fail. The disruption, which extended well into Friday, impacted a suite of essential tools including Outlook for email, Teams for communication,
and even the AI-powered Copilot. Reports from outage trackers showed tens of thousands of initial complaints, which quickly ballooned as the issue cascaded across 74 countries. The effects were felt far and wide, disrupting government agencies in Japan, Australia, and the US, halting court proceedings in the Philippines, and impacting airlines in India and shipping companies in Denmark. Some organizations were even forced to revert to older technologies like fax machines to maintain critical communications.
The Technical Explanation
In the immediate aftermath, speculation about a potential cyberattack was rampant, but Microsoft quickly stated it had found no evidence of malicious activity. The root cause was more mundane, yet just as crippling: a server configuration issue. According to Microsoft's updates, the problem began when a large number of its servers unexpectedly became unavailable. This wasn't due to a spike in user traffic but rather an internal failure. The remaining active servers were instantly overwhelmed, unable to handle the redistributed load. The company later specified that a process responsible for establishing secure connections was becoming overloaded, causing servers to stop responding to requests entirely. This created a domino effect where even servers that were restarted would fail again once traffic was directed back to them. The issue appeared to be linked to a recent network configuration change, which the company began to roll back in an effort to stabilize the system.
The Difficult Path to Recovery
Fixing an issue of this magnitude across a global network is not as simple as flipping a switch. Microsoft's engineers worked through the night, but initial attempts to fix the problem did not hold. The company reported that some recovery actions were 'not taking as expected', indicating the complexity of the problem. By Friday morning, Microsoft reported 'incremental improvement' and 'early signs of stabilization' as they worked to apply software patches and bring more virtual machine capacity online. However, services remained inconsistent, with many users still experiencing intermittent access or significant email delays. To prevent data loss, Microsoft extended its email timeout window to 48 hours, a measure to protect messages that were stuck in processing queues during the recovery.
A Reminder of Cloud Fragility
This massive outage serves as a stark reminder of the modern world's deep reliance on a small number of technology giants for critical infrastructure. While cloud services offer immense power and convenience, they also introduce a single point of failure that can have global consequences. The incident highlighted that organisations using Microsoft's cloud-based Exchange Online were the most severely affected, while those with on-premises servers were somewhat spared. The disruption went beyond just email and meetings, impacting AI tools from OpenAI and Anthropic that rely on Microsoft connectors, and even halting financial transactions in Europe. As businesses and governments continue to migrate more of their essential functions to the cloud, this event underscores the crucial need for resilience planning and understanding the interconnected nature of the digital ecosystem.
