Microsoft's 24-Hour Outage Traced to July Bug, Traffic Shift, and Capacity Shortfall
A preliminary incident report from Microsoft reveals that a dormant defect introduced in July, combined with network maintenance and reduced server headroom, triggered one of the company's largest outages, affecting Microsoft 365 and Exchange for nearly a full day.

Microsoft has released a preliminary Post-Incident Report detailing the root causes behind a major outage that struck its cloud and email services last week. The company attributes the disruption to a convergence of three separate technical problems: a latent software defect that had persisted since July, a shift in network traffic patterns during routine maintenance, and degraded infrastructure capacity. The incident, which lasted from November 12 at 16:20 UTC through November 13 at 16:10 UTC—just under 24 hours—disabled significant portions of Microsoft 365 and Exchange infrastructure. Microsoft classified the event as "a technical issue not related to cyber or AI/agents."
The outage swept across businesses and government agencies worldwide that rely on Microsoft's cloud services and email systems. The company has assigned incident IDs EX1550543 and MO1550821 to track the event. While the full preliminary report requires Microsoft 365 administrator credentials to access on the company's status pages, Office Watch has published the report, and the University of Pennsylvania maintains an archive of Microsoft's service-health notifications from the period.
A perfect storm
The first contributing factor was a defect introduced in a July deployment to Microsoft's cloud gateway. Under specific circumstances, this bug could exhaust CPU and memory resources, rendering systems unresponsive. Microsoft states the flaw went undetected during testing because the pre-release hardware lacked the processing power of production systems already in operation, where higher CPU core counts on newer machines apparently triggered the resource exhaustion.
The second factor involved server load and traffic patterns. According to Microsoft's account, "Traffic shifts during standard network maintenance and upgrades exposed a latent bug." The company maintains two separate gateway systems to provide redundancy; if one fails, the other handles incoming traffic to keep services running. About 15 hours before the outage began, traffic from Teams was rerouted from the backup gateway to the primary gateway, which manages Exchange Online access. This standard failsafe procedure meant the primary gateway suddenly carried higher traffic volumes with a different composition of requests than before. As demand accumulated, the affected servers reached capacity and began rejecting new connections.
The third factor was insufficient infrastructure capacity. An unrelated fault had left Microsoft's gateway server fleet operating at 14 percent below its intended performance level. Under normal circumstances, this missing capacity buffer would have absorbed the increased demand, but when combined with the traffic redistribution and the underlying defect, it created lock contention as requests continued to mount.
As business activity in the United States reached peak hours, the gateway servers began to freeze. Failover systems designed to redirect traffic away from unresponsive servers encountered the same resource exhaustion on the supposedly functioning backup gateways, creating a cascading failure. More than 80 percent of Microsoft's affected gateway servers became locked in the incident. Regional isolation mechanisms that should have contained the problem appear to have malfunctioned, though Microsoft's current report does not explain why these safeguards failed.
No easy way out
Restoring service proved difficult because the underlying defect required a targeted software update from Microsoft's engineering teams. Simply restarting the gateway servers proved ineffective; when servers returned to operation, massive queues of pending requests immediately resumed processing, triggering the same resource exhaustion that had caused the original failure. While Microsoft could redirect services such as Teams and OneDrive away from the compromised infrastructure, scaling constraints prevented Exchange Online from being rerouted in the same manner.
Rolling back to previous software versions offered no solution either, since the defect had existed since July and was present in all recent versions. Technicians instead patched servers individually and restored them to service in stages. Microsoft has stated it will implement additional server capacity and traffic-management controls to prevent recurrence. The company also adjusted email delivery timeouts on affected systems to prevent delayed messages from being rejected as undeliverable.
Microsoft plans to publish a complete final incident report on November 20, which may offer more detail about the failure of regional isolation systems and outline additional remediation steps. By Monday morning, all Microsoft cloud services had returned to normal operation.