Hidden software flaw, traffic surge and reduced capacity behind Microsoft 365 outage; over 80% of affected servers hit
New Delhi: Microsoft has attributed its recent global Microsoft 365 outage to a combination of a hidden software defect, a sharp rise in traffic and reduced server capacity, saying the three factors together overwhelmed systems and disrupted services including Teams, OneDrive, SharePoint and Office applications.
In an update shared on X, Microsoft’s Executive Vice President for Copilot, Agents and Platform apologised to customers and outlined early findings from the company’s investigation into the November 12 disruption.
According to the company, the problem began in a networking system that sits in front of many Microsoft 365 services. A previously undetected software flaw caused certain servers to become overloaded and stop responding under specific conditions.
Microsoft said the code change associated with the flaw was first deployed in July and had operated globally for more than three months without any apparent issue. The defect went unnoticed because the company’s testing environment used traffic patterns and hardware that differed from those in its live production systems.
The company said traffic on the affected system rose sharply after Teams traffic was shifted from a backup system to the primary system about 15 hours before the outage. At the same time, nearly 14 per cent of server capacity was unavailable due to an unrelated issue from the previous day.
While neither factor alone would have caused an outage, Microsoft said their combination exposed the hidden flaw. Requests then accumulated until servers stopped responding during peak traffic hours in the United States.
Microsoft said its automated systems initially attempted to redirect users to healthy servers. However, the additional traffic increased the likelihood that the same failure would occur elsewhere, triggering a rapid chain reaction across the system.
At the peak of the incident, more than 80 per cent of the affected servers had stopped responding, according to the executive.
The company also said a separate issue prevented its internal safety systems—designed to contain failures within individual regions—from functioning properly, allowing the disruption to spread globally.
Recovery proved difficult because servers that came back online were immediately hit by a backlog of pending traffic and stopped responding again. Microsoft said its existing automated recovery mechanisms were not designed to restore services after such a large-scale failure.
Engineers eventually identified the faulty code and developed an update that restored normal operations.
Microsoft said it will publish a full review later this week with additional details and corrective measures. The company had earlier said the incident was a technical issue rather than a cyberattack or an issue involving AI agents.
(Inputs from IANS)