Microsoft Final Report Blames Latent Bug and Testing Gaps for Global Outlook Outage
A latent software bug, introduced in July but left dormant for months, caused a global Microsoft 365 outage on November 12 and 13 by overwhelming newer, high-performance gateway servers. Microsoft confirmed the service disruption was not caused by a cyberattack or artificial intelligence, pledging 13 technical fixes to address testing gaps by March 2027.
Testing Gaps on Newer Hardware
The bug triggered a traffic jam within the Windows process responsible for establishing incoming connections on network gateway servers. Because these newer servers feature more CPU cores and memory, they attempted to process more simultaneous requests than their predecessors, causing the system to exhaust its computing resources, according to the final report released Friday, as reported by MLQ.ai.
Microsoft acknowledged that its pre-deployment testing failed to account for the specific hardware and traffic mix encountered during the incident. “Our testing did not adequately represent the combination of hardware and traffic mix encountered during this incident,” the company said in its final report, as noted by Geekwire.com.
To prevent a recurrence, Microsoft plans to update its testing environments to better mirror the production hardware currently running its services, Journalarta.com reported.

Three Conditions Triggered Failure
While the bug had existed since July, three concurrent conditions on November 12 forced the system into a state of failure. Nearly 14% of gateway capacity was offline for scheduled maintenance, and a routing glitch in the US East region had shifted all global Teams traffic onto the email gateways roughly 16 hours earlier, MLQ.ai reported.
As North American traffic approached its daily peak, the remaining servers buckled under the combined load.
The company’s preliminary report had previously attributed the loss of capacity to an unrelated fault, a claim corrected in the final document, Geekwire.com reported. This influx of traffic overwhelmed the gateways, leading to a cascade of errors across services including Outlook, Exchange Online, Teams, Copilot, SharePoint, Defender, and Purview.
Cascading Safeguards Spread Outage
The failure was compounded by automated traffic management systems that attempted to mitigate the load. As servers became unresponsive, the system redirected traffic to healthy nodes, which then suffered from the same latent bug and subsequently failed, Journalarta.com reported. At the lowest point of the outage, more than 80% of North American gateway servers were unresponsive, forcing the system to shift traffic to other global regions.
Recovery efforts were initially delayed because servers returning to service were immediately overwhelmed by the backlog of waiting clients, causing them to freeze again, MLQ.ai reported. Microsoft plans to implement a more gradual traffic restoration process and will change how traffic shifts between gateways to prevent future cascading failures, according to Journalarta.com.
I want to share what we know so far. This was not a cyberattack or an issue caused by AI tooling. Our amazing team was able to work through a very challenging situation to restore service. Our investigation points to technical challenges and failure modes (cascading failures +… https://t.co/lFQu0CgOQG
| Factor | Impact on Gateway Servers |
|---|---|
| Planned Maintenance | 14% capacity offline |
| Teams Routing Glitch | Global Teams traffic shifted to email gateways |
| North American Traffic | Climbing to daily peak capacity |
Infrastructure Flaws Triggered Global Cascades
The outage began at 4:20 p.m. UTC on November 12 and persisted until 4:10 p.m. UTC the following day, as reported by MLQ.ai.
As these gateway servers became unresponsive, Microsoft’s automated traffic management systems attempted to contain the incident by shifting workloads to nearby healthy nodes.

Microsoft Commits to Infrastructure Overhaul
These measures include separating email traffic from other service surges, modifying how traffic shifts between regions, and implementing a more gradual restoration process for recovering servers. The company also plans to improve its incident response by shortening customer notification intervals, according to Journalarta.com.
CNBC highlighted that these changes address the broader challenge of maintaining stability amid rapid software deployment cycles.
Resilience and Future Safeguards
Microsoft has committed to 13 specific improvements to its infrastructure to be completed by March 2027, including separating email traffic from other service surges and shortening customer notification intervals, Journalarta.com reported. Some of these updates are scheduled for implementation by the end of 2026, though others currently lack a specific completion date.
The company confirmed that no business email was lost during the disruption, though some delayed messages sent to consumer mailboxes that had been inactive for over 30 days bounced back to senders, Journalarta.com reported.
Addressing the impact of the outage during the Ignite 2026 conference in San Francisco, Microsoft Chairman and CEO Satya Nadella offered an apology to customers.
“I know we have work to do to provide that resilience that you can count on,” said Satya Nadella, Microsoft Chairman and CEO, as reported by Geekwire.com.