For WindowsForum readers the main story is Microsoft's. The Microsoft 365 outage was the longest of the group, and Microsoft's preliminary post-incident report gives far more detail than any other company has published so far.
The Microsoft 365 outage: almost a full day with limited email
According to Microsoft's preliminary Post-Incident Report, dated November 15, the outage began at 4:20 PM UTC on November 12 and ended at 4:10 PM UTC on November 13, lasting nearly 24 hours. It is tracked as EX1550543 (Exchange Online) and MO1550821 (Microsoft 365). Exchange Online took the worst of it, with many users unable to access their mailboxes or send and receive email normally.
Other services were affected too. The report names Exchange, Copilot, Teams, SharePoint, OneDrive, Defender, Purview, Fabric, Intune, Planner, Power Apps, To Do, Power BI and Universal Print. That list overstates the damage. According to 365Fanatic's summary of the report, services that use Exchange Online mostly stayed online, but the parts that work with Exchange were affected. Most of the suite stayed up. The parts that depend on Exchange Online stopped working when it did.
Three problems at once
Microsoft's report does not point to a single error. It describes three problems that happened together:
- A bug that had been sitting there since July. The report blames a previously undetected defect introduced through a Windows update deployed to its Microsoft 365 networking gateway servers in July 2026. These servers are Microsoft's front door: connections to Exchange Online and many other Microsoft 365 services pass through these servers. WindowsForum's earlier coverage of the report described it as a lock-contention bug.
- Extra traffic from Teams. About fifteen hours before the outage, Teams automatically failed over from its alternative gateway to the primary one, which also serves Exchange Online. That happened at 1:07 AM UTC on November 12.
- Isolation that failed. Microsoft's partitions between components and regions should have kept the failure contained. Because of a separate issue they did not, and the outage spread globally.
That explains why Teams users and Outlook users had very different days. Teams, OneDrive, SharePoint and the Office apps could fail out of the faulty gateway and so recovered earlier (Teams chats and meetings worked without issue for most of the outage). Exchange Online cannot do that for scale reasons. Messageware drew the lesson: Exchange Online relied on a single gateway system, so when that gateway failed, email had nowhere to go. Teams' automatic failover helped Teams and added load to the gateway Exchange depended on.
The timeline also includes a patch. WindowsForum's breakdown of the report lists it as deployed at 07:20 UTC on November 13, but the impact window did not close until that afternoon.
Does this affect your Windows machines?
"Windows defect" will make some admins check their own patch history, so here is the important point. According to the report, this defect was the result of a fix that was prepared specifically for Microsoft 365 networking gateway servers; neither the underlying defect nor the fix applies to broader Windows deployments. Your desktops, servers and July cumulative updates are not affected by this bug.
Not a cyberattack, though attackers took advantage
Microsoft reiterated that the outage was a technical failure, with no indication of malicious activity or involvement from AI agents. That did not stop criminals from using it. Messageware reports that attackers moved quickly to exploit the confusion, sending fake outage notices that warned staff their "held" email would be deleted unless they clicked a link. On November 13, Microsoft Threat Intelligence found a phishing campaign aimed at organizations across several US industries.
An outage cannot steal credentials. A convincing fake outage notice sent while people are confused can.
What admins should take from this:
- Microsoft said it was not a cyberattack and that you do not need to change anything in your tenant. Rebuilding Outlook profiles or resetting passwords would not have fixed a server-side outage.
- Get incident status from the Service health page in the Microsoft 365 admin center, which shows the details for your organization. Do not rely on emails claiming to report it.
- Send users a reminder now. Genuine Microsoft outage communications do not ask them to click a link to rescue quarantined mail.
- Watch for the final report. Microsoft plans to publish its final Post-Incident Report by November 20, 2026. The preliminary report is explicitly not the final word, and Microsoft's explanation already changed once during the incident.
This is not a one-off. Messageware counts it as Microsoft's third major Microsoft 365 outage of 2026, after incidents in January and September. That is a fair point to raise when you review your SLA.
Section summary: A bug introduced in July, extra traffic from a Teams failover and partitions that failed to contain the problem knocked out Exchange Online worldwide for nearly 24 hours. The defect does not affect customer Windows systems. Microsoft says no attacker was involved, but phishers used the outage as cover.
Apple: iCloud sign-in problems on November 17
A few days later, Apple had its own outage. TechCrunch reported that Apple's System Status page showed iCloud Account & Sign In problems affecting all users, with service possibly slow or unavailable. The App Store, AppleCare on Device and Apple Music also had problems. Because the sign-in service was down, iPhones could prompt users for their Apple ID password again, other Apple-account logins could fail, and some Reddit users said they could not activate new iPhones.
When DigitalToday published, Apple had acknowledged the problem on its status page but had not given an explanation. TechCrunch later updated its story with an Apple statement, first given to Bloomberg, blaming an internal update and saying it was not a cyberattack. Apple said many systems were back up and it was restoring the rest. TechCrunch also reported that the status page added Apple School Manager, iCloud Mail, iCloud.com web apps, Schoolwork and the Volume Purchase Program to the affected list. The last two matter for schools and organizations that manage Apple devices.
Banks, Workday and the NSA
The rest of the cluster is less well documented:
- Banks (November 16): DigitalToday reported online banking access problems for customers of Chase, Bank of America, Wells Fargo and Capital One. TechCrunch described a brief outage of Chase's banking app that was blamed on a software glitch. Nobody has shown that every bank on the list failed for the same reason. Reports of customers being unable to log in are not the same as a confirmed shared cause.
- Workday: DigitalToday reported a large outage at the HR and payroll provider that left some customer employees unable to reach their accounts. No cause, duration or scope has been independently confirmed, so treat it as a reported incident only.
- NSA: DigitalToday and TechCrunch both report an outage of the agency's internal classified network, which TechCrunch said was still partially down at its time of writing and which both outlets said was attributed to software problems. Public detail on the NSA is thin, as you would expect, so do not count this as a fully explained case.
The FAA ground stop
The incident that drew the most international attention was the FAA's order on Friday, November 13, halting domestic departures across the U.S. DigitalToday reports that the order came after safety concerns over a communications failure. Transportation Secretary Sean Duffy blamed outdated air traffic control technology and overloaded routers. Markwayne Mullin, identified as Homeland Security Secretary, said the problem happened while the FAA was updating an aging system, was not caused by threats against the United States, and was not related to the Microsoft outage. The investigation was still under way, so this account could change.
What the incidents have in common
Do these outages share a cause? Not as far as anyone has shown. No source links Microsoft, Apple, the banks, Workday, the NSA and the FAA to a common cause or a coordinated campaign. They happened close together, and that may be all.
They do show a common pattern. Microsoft and Apple blamed their own internal changes. The FAA explanation also involves an upgrade that went wrong. In Microsoft's case, the automatic failover meant to improve resilience moved load onto the system that eventually collapsed. That matches DigitalToday's suggestion that pursuing efficiency may be costing stability. This is my analysis, but it is common industry knowledge that the more tightly services are coupled, the harder it is to test failover paths realistically, and that is where hidden defects tend to stay hidden.
There are reasonable counterpoints. Very large platforms will always have some outages. A week with several of them does not prove reliability is getting worse overall. "Not a cyberattack" also only applies to the incident in question. It does not mean the threat has gone away, as the phishing campaign during the Microsoft outage showed.
Bottom line for Windows and Microsoft 365 shops
- The Microsoft 365 outage (EX1550543 / MO1550821) ran from November 12 to 13. Microsoft's preliminary report blames a bug on its gateway servers, a Teams failover and partitions that did not contain the failure. It was not an attack.
- The defect affects only Microsoft 365's gateway servers, not ordinary Windows installations, and no tenant changes are needed.
- Expect the final report by November 20 and compare it with the preliminary one.
- Assume the next outage will come with phishing that imitates it, and prepare users before it happens.
- If your business continuity plan assumes Teams and Outlook fail together, update it. This outage showed they can fail separately.
There is not one cause behind all of these outages. Some were caused by internal updates or bugs. Others have not been fully explained yet.
Correction (November 17, 2026): An earlier version of this article said the FAA's nationwide ground stop was ordered "the previous weekend." The ground stop was ordered on Friday, November 13, 2026, and ran from 3:54 p.m. to 8:15 p.m. ET; the resulting flight delays stretched into the weekend.
Update: NSA attributes network degradation to routine maintenance (November 18, 2026)
The NSA said its network experienced a “temporary degradation” during routine maintenance on November 17, according to Reuters reporting published by The Business Times. The agency said its mission-critical capabilities remained fully operational. That adds an official explanation to the earlier reports of software problems, though it does not establish the full scope or duration of the disruption.
Reuters also reported that JPMorgan Chase attributed problems processing transactions to a software glitch. It described reports that other banks had gone offline as erroneous. That is narrower than the earlier reports of access problems at several banks; it does not establish a shared banking-system failure.
White House spokeswoman Lauren Bis said there was no evidence that the disruptions were caused by a cyberattack, according to Reuters. The report also found no evidence that job cuts at the affected organizations caused the outages.
For IT teams, the new statements provide more detail about two incidents but do not identify a common cause or call for changes to Windows systems or Microsoft 365 tenants.
References
- Software flaws, not hacks, drive wave of major IT outages - 디지털투데이 디지털투데이 · 2026-11-17T20:58:19+00:00
- As string of tech outages raises alarm, Trump says everything is ‘fine’ - The Business Times The Business Times · 2026-11-17T22:58:34+00:00
- Microsoft 365 outage 12 November 2026: post-incident report 365fanatic.com