Home / Blog / Inside Microsoft’s giant Outlook outage: At 9AM last…
Tech News

Inside Microsoft’s giant Outlook outage: At 9AM last Thursday

A dormant July networking bug and a Teams failover triggered a 24-hour Microsoft Outlook outage that knocked over 80 percent of Exchange servers offline.

By Dillip Chowdary • Nov 18, 2026 • Source: The Verge

Inside Microsoft’s giant Outlook outage: At 9AM last Thursday
Image: The Verge

At a glance

Product
Microsoft Outlook and Exchange Online
Incident severity
Sev0 catastrophic service disruption
Outage duration
Nearly 24 hours
Server impact
Over 80 percent of Exchange servers unresponsive
Root trigger
Latent Windows lock contention in gateway servers
Bridge call
Hit maximum limit of 1,000 engineer participants

Microsoft experienced one of its worst outages in history as a cascading failure disabled Exchange Online and cut off Outlook email for millions of enterprise, commercial, and government users worldwide for nearly 24 hours. The disruption began on a Thursday morning when mail servers fell offline, according to details disclosed in The Verge's report. Automated failover systems rapidly amplified the incident until more than 80 percent of Microsoft's global Exchange server infrastructure became completely unresponsive.

This report details the technical sequence behind the failure, examining how traffic failovers, latent gateway defects, and reduced capacity triggered a major cloud outage. It is written for systems administrators, engineering leaders, and IT professionals responsible for enterprise communications reliability. The analysis traces the event from initial European server faults through troubleshooting obstacles and planned testing reforms.

01Failover cascade across global Exchange servers

The disruption originated at Exchange Online facilities in Switzerland, where servers fell offline shortly after 9 AM on a Thursday. Built-in failover routines attempted to redirect live traffic to alternative regions, knocking systems offline in Mexico. The redirected volume rapidly overwhelmed downstream clusters across multiple countries instead of isolating the failure.

Within two hours, more than 80 percent of Microsoft's global Exchange server infrastructure had become unresponsive. Microsoft escalated the event to sev0 in its internal incident tracking system, designating a catastrophic level of service disruption. Hundreds of engineers immediately mobilized to find the root cause and restore messaging services across the fleet.

Under normal conditions, automated failover systems route traffic toward healthy servers to maintain continuous email access. During this incident, however, the automated redistribution propagated instability across the global network topology. As successive regions buckled under the redirected load, the failover architecture directly accelerated the collapse of the broader platform.

02Network boot loops and attack investigations

Responding engineers analyzed crash reports, trace logs, and kernel dumps while servers entered maintenance mode under unusually high CPU loads. Technicians attempted to restart failed machines, but returning nodes were immediately overwhelmed by waiting client traffic queues. The incoming connection spikes caused servers to crash repeatedly before completing their initialization sequences.

The recovery effort was further hindered by the network booting architecture used across Exchange servers. Machines boot and run their operating system over the network using Preboot Execution Environment protocols to permit centralized remote management. When hundreds of machines rebooted simultaneously, the burst of traffic overwhelmed monitoring systems, creating an ongoing reboot loop.

Standard change-reversion procedures failed to restore stability, raising external speculation regarding a cyberattack or AI coding errors. Microsoft engineers investigated potential attack vectors and coordinated with DDoS specialists to determine whether malicious activity was occurring. Crash dumps and internal telemetry soon eliminated outside intervention, confirming an internal defect had destabilized the infrastructure.

03Latent gateway bug and multicore hardware

Advertisement

Tech Pulse Daily

Get tomorrow's pulse first

Join engineers who read Tech Pulse before stand-up. Free, weekday mornings.

After ruling out external attacks, engineers examined the networking stack and discovered an unusual bug on Microsoft 365 networking gateway servers. The defect had resided quietly on the gateway fleet since July without causing disruptions. Charles Lamanna, leader of the Copilot, Agents, and Platform team, stated that a latent defect caused servers to become overwhelmed and unresponsive under certain load conditions.

The engineering groups supporting Office, Outlook, and Teams were placed under Lamanna following the departure of former division head Ryan Roslansky. A preliminary post-incident report described the flaw as affecting select Windows servers due to lock contention inside an internal Windows process. The code fix was implemented specifically for Microsoft 365 gateway machines and did not affect standard Windows editions.

Microsoft admitted that internal testing failed to uncover the issue earlier because of hardware differences in its evaluation labs. Lamanna noted that lock contention worsened on production machines featuring more cores and newer hardware architectures. Microsoft's automated load testing had primarily focused on weaker hardware with fewer processor cores, allowing the dormant concurrency flaw to escape detection.

04Teams gateway migration and reduced capacity

The latent defect required two independent operational events to expose its vulnerability across the production infrastructure. Fifteen hours before the Outlook outage began, Microsoft Teams traffic failed over from an alternative networking gateway to the primary gateway supporting Exchange Online. This planned failover altered normal traffic patterns and concentrated substantial connection loads directly onto the shared gateway infrastructure.

Simultaneously, approximately 14 percent of the main gateway's server capacity was offline due to an unrelated technical issue from the previous day. Under standard conditions, the network gateway maintained sufficient headroom to absorb the additional Teams connection volume without failure. However, operating with reduced server capacity during high-demand windows eliminated safety margins across the remaining active nodes.

This combination of factors produced a breaking point across the gateway infrastructure. The arrival of redirected Teams traffic, reduced capacity, and the dormant July bug triggered the lock contention flaw on the gateway servers. The resulting bottleneck prevented servers from processing requests, causing Exchange Online to cascade offline across global tenant environments.

05Thousand engineer bridge and traffic surges

Restoring Exchange Online required nearly 24 hours of continuous troubleshooting across Microsoft's engineering organization. The company maintained an active incident bridge call that reached its maximum limit of 1,000 participants as personnel coordinated remediation across product divisions. Recovery operations were complicated by overlapping team interventions as multiple groups attempted concurrent fixes during peak United States traffic hours.

Recovery proceeded slowly because automated recovery processes were not designed to restore services from such an extensive unresponsive state. As servers returned to service, waiting client traffic immediately swamped them, forcing the hardware back into unresponsive conditions. Standard rollbacks, flighting routines, and feature flags were incapable of resolving the issue because the defect had been deployed months earlier in July.

Microsoft is preparing a final post-incident review for publication later this week detailing permanent architectural remedies. The incident highlights the necessity of expanding testing environments to mirror production hardware with higher core counts. Strengthening test coverage and automated recovery controls will be central to ensuring future gateway strains do not trigger catastrophic global disruptions.

Frequently asked questions

What caused the Microsoft Outlook outage?

A latent July bug in Microsoft 365 gateway servers caused lock contention inside an internal Windows process under high load. This was triggered when Teams traffic failed over to the gateway while 14 percent of its capacity was offline.

How long did the Exchange Online service disruption last?

The disruption lasted nearly 24 hours before engineers stabilized servers. At its peak, over 80 percent of Exchange servers were unresponsive.

Did a cyberattack or DDoS campaign take down Outlook?

No, Microsoft investigated attack vectors with DDoS specialists, but crash logs quickly ruled out foul play. The failure stemmed entirely from an internal software defect and compounding traffic shifts.

Why did automated recovery systems fail to restore service?

Returning servers were immediately swamped by waiting traffic queues, sending them back into an unresponsive state. Standard rollbacks and feature flags were also ineffective against the dormant July bug.

Sources

Dillip Chowdary

Author

Dillip Chowdary

Writes Tech Bytes coverage of AI, engineering, and the tools that actually ship. Editor of Tech Pulse Daily.

Related on Tech Bytes

Advertisement

5-min tech signal

Weekday briefing for engineers who skip the noise.

No spam · Unsubscribe anytime

Advertisement

✈️ CareerPilot

Your AI job-search copilot

Match your resume against live Ashby, Greenhouse & Lever openings — fit scores, job-specific resume optimization and email alerts.

Find matching jobs →

Free Tools

Browse all tools →