- Microsoft (NASDAQ:MSFT) says a software defect present since July triggered last week's worldwide disruption, which began with traffic rerouting on November 11 and peaked with over 80% of servers in an affected gateway unresponsive.
- The main outage lasted nearly 24 hours on November 12–13; Microsoft took about 30 hours to identify the broader problem and deploy a fix.
- A final incident report is promised by November 20, with questions remaining over why regional isolation failed and how testing missed the flaw.
A Latent Defect Exposed
Microsoft has attributed last week's widespread disruption to a software bug that went undetected for months. According to the company's preliminary investigation, the defect was introduced in July and lay dormant until a routine traffic rerouting event on November 11 exposed it, causing cascading failures across Microsoft 365 services including Outlook, Teams, and Copilot.
At the peak of the outage, more than 80% of servers in the affected Microsoft 365 networking gateway were unresponsive, according to Microsoft. The figure refers to that specific gateway, not all Outlook servers. The main worldwide outage lasted almost 24 hours on November 12–13, and it took Microsoft approximately 30 hours to identify the broader problem and begin deploying a fix.
The company said the failure was not the result of a cyberattack or AI malfunction. Instead, three interacting problems created the perfect storm: the July software change caused internal tasks to block one another under particular workloads, eventually exhausting server resources; Teams traffic was moved onto the same gateway that supported Exchange Online, changing its workload; and nearly 14% of the gateway's capacity was already unavailable due to a separate issue.
Cascading Failures and Recovery
As servers failed, automated systems redirected requests to healthy servers, triggering the same defect there. Returning servers to service did not initially help: queued requests overwhelmed them again. Exchange Online—the email platform behind Outlook—suffered most because it could not move away from the gateway at the required scale, while Teams, OneDrive, SharePoint, and Office applications could partly bypass it.
Engineers ultimately deployed an update reverting the faulty code, restarting groups of servers in stages. Microsoft has since added capacity, monitoring, and limits on traffic redistribution. It also extended the email-delivery window so queued messages could be delivered rather than returned as undeliverable.
An important finding concerns testing: Microsoft's load tests tended to use weaker machines with fewer processor cores, while the defect was worse on newer, higher-core-count hardware. The preliminary report leaves unanswered why regional isolation barriers failed to contain the cascade.
Business Impact and Next Steps
The disruption affected employees and customers worldwide, with more than 130,000 Outlook problem reports at the peak, spanning approximately 72–74 countries or territories. Public-service users in South Africa faced communication interruptions, and hospitals and banks experienced impaired communications, according to reporting on the outage's wider effects.
The incident raises fresh concerns over cloud infrastructure resilience and enterprise dependence on Microsoft. The company's own resilience guidance acknowledges that concentrating critical services with one provider can increase the impact of a failure.
Microsoft's latest quarterly results, released October 28 for the quarter ended September 30, showed revenue of $92.3 billion, up 19% year-over-year, with Microsoft Cloud revenue of $63.9 billion. The outage occurred after that period and is not reflected in those results.
The next major milestone is Microsoft's promised final incident report by November 20. Customers and regulators will be watching for answers on why regional containment failed, how testing will cover newer hardware, and how recovery systems will avoid overwhelming servers as they return. The incident is likely to intensify scrutiny of shared dependencies and continuity plans, particularly for regulated industries under frameworks like the EU's Digital Operational Resilience Act.
Microsoft did not immediately respond to a request for comment on the final report's scope.