Lessons from Microsoft's November 2026 Patch Tuesday

Patch Tuesday took a month off from breaking records, and Outlook took the spotlight instead. Of the 1,094 vulnerabilities Microsoft fixed on November 10, attackers had already exploited one: a Moderate-rated Outlook bug that quietly hands a user’s password hash to whoever sent the file they opened. Two days later, Outlook stopped connecting for users around the world, and help desks spent the afternoon fielding the same question: did the patch do this?

Microsoft's preliminary findings have traced the outage to its own network, but the week left admins with two Outlook problems to explain to the business. Somewhere in your organization, a user may already have opened the wrong message file, and the update you deploy this week won’t get their hash back. Meanwhile, the outage was a reminder of what a single change can do to a service everyone depends on.

Lesson 1: Patching Outlook won't undo stolen hashes  

Microsoft confirmed that attackers exploited CVE-2026-124194 (CVSS 6.5) before the fix shipped, and the Cybersecurity and Infrastructure Security Agency (CISA) added it to its Known Exploited Vulnerabilities (KEV) catalog with a November 24 deadline for federal agencies. The flaw leaks the user’s NT LAN Manager version 2 (NTLMv2) hash when they open a crafted message file, and some of those hashes may already be gone. The Preview Pane isn’t an attack vector, so exposure is limited to users who actually opened a file. An attacker can crack a stolen hash offline or relay it to sign in as that user wherever signing isn't enforced. When that user approves payments or manages identities, the exposure becomes account takeover, payment fraud, or a foothold in the domain, and it stays open on a fully patched machine.

For this CVE, close the ticket when the affected passwords are reset. Search firewall and endpoint logs for Outlook connecting to external addresses over Server Message Block (SMB, TCP 445), and count blocked attempts too, because each one likely marks a user who opened a malicious file. Rotate credentials for administrators first, then for finance and identity teams, and report progress to leadership as users exposed against users rotated.

The outage made the attacker’s job easier. Users who watched Outlook fail for an afternoon are primed to trust a “resend” or “mailbox resync” email, and an attached message file is exactly how CVE-2026-124194 gets delivered, so a short warning to users costs little.

If you haven’t moved privileged accounts into the Protected Users group or started auditing NTLM use ahead of Microsoft disabling it by default in the next major Windows Server release, November is a good time to start.

Lesson 2: The outage shows why every change needs a rollback plan   

On November 12, Exchange Online servers began failing automated health checks and dropping out of service, which pushed connection load onto the rest of the infrastructure until Outlook, Teams, SharePoint, and other Microsoft 365 services stalled across regions (incident EX1550543 in the Microsoft 365 admin center). Microsoft found no evidence of malicious activity. Its preliminary findings trace the failure to a problem introduced during an earlier network capacity upgrade that only surfaced after a later network change. To recover, Microsoft paused all deployments and configuration changes and rolled back the recent network change. It also held delayed mail for 48 hours, so most messages sent during the outage arrived once service recovered.

Two parts of that story apply directly to patching. The failure was set up by one change and triggered by another, and patches behave the same way: a fix that passed your pilot can still break when the next update or configuration change lands, so keep a change log you can line up against an incident timeline. And Microsoft’s first move was to freeze changes and roll back. If your team can’t do both within hours for the Office update you push this month, your pilot is the place to find out.

The outage also tested patch discipline from the other direction. A user whose Outlook stops connecting the day after you push a patch will blame the patch. Check Microsoft 365 service health before you roll anything back, because reverting the Outlook fix during a service-side outage reopens the only flaw attackers are known to be using from November's Microsoft release. For leadership, it’s a useful prompt to ask how the business runs when email is down for an afternoon, and who decides whether to pause your own changes while a provider recovers.

Lesson 3: The Outlook fix skips Windows Update

The Outlook fix ships through Office's update channel, so a device can show the latest Windows cumulative update and still run a vulnerable Outlook. Microsoft lists Microsoft 365 Apps and both Office Long-Term Servicing Channel (LTSC) releases, 2021 and 2024, as affected.

The business cost here is false assurance. A dashboard that reads 98% compliant because it only counts Windows updates gives leadership and auditors a number that doesn’t match the exposure. Track compliance per update channel, and count a host as patched only when every update it needs is installed. A patch management tool that reports every update channel side by side makes this the default view. ManageEngine Patch Manager Plus, for example, deploys Microsoft 365 updates through the same Automated Patch Deployment (APD) tasks as Windows updates, with Click-to-Run support from on-premises build 10.1.2137.1, so one compliance report covers both.

On-premises Exchange, which the Exchange Online outage didn’t touch, has a two-channel problem of its own. CVE-2026-116928 (CVSS 9.8) is a Windows Search flaw an attacker triggers by emailing a crafted attachment that Exchange indexes automatically. Its fix ships in the Windows Server cumulative update, separate from the Exchange Server Subscription Edition (SE) security update, so each Exchange host needs both.

Lesson 4: Patch exploited flaws first, with leadership sign-off  

CVE-2026-124194 sits behind all 63 Critical flaws in a severity-sorted queue, yet it was the only one under attack. At more than 1,000 fixes a month, you can't patch everything inside one change window, so the order you patch in is a business decision worth making once, in writing, with leadership's sign-off. A workable rule says known exploitation outranks everything, and among unexploited flaws, exposure to the internet or to identity systems outranks the CVSS score. CISA’s KEV deadlines give you an outside clock: a two-week window for an exploited flaw is easy to defend when you request an emergency change.

Under that rule, November’s order is clear: Outlook first, then domain controllers for an unauthenticated Lightweight Directory Access Protocol (LDAP) remote code execution (RCE) flaw (CVE-2026-119958, CVSS 9.8), then Exchange hosts for CVE-2026-116928 (CVSS 9.8). For the patch-now path, shrink the pilot to a few hours on a representative group and keep the rollback plan from Lesson 2 ready.  

Also on November 10   

.NET 8, .NET 9, PowerShell 7.4 long-term support (LTS), and PowerShell 7.5 stopped receiving fixes, and Windows 11 version 23H2 Enterprise and Education received their final security update. None of those runtimes shows up as a missing Windows update, so inventory them through software discovery and put a costed migration plan to .NET 10 and PowerShell 7.6 in front of leadership.

What November adds to the mix

September's lessons dealt with volume and October's with which machines a fix matters for. November's are about what a clean install still leaves behind, and about the changes, Microsoft’s and your own, that a week like this puts under the microscope. See how Patch Manager Plus deploys and reports Office and Windows updates from one console before the next Patch Tuesday on December 8.

This week, close the exploited gap first: push the Outlook fix through Office’s update channel and confirm it actually landed, because the Windows report won’t show it. Count it as fixed only once the exposed users have reset their passwords, and share the users-exposed-against-users-rotated number with leadership. Warn users about outage-themed attachments while they’re still primed to trust them, and check Microsoft 365 service health before anyone rolls the fix back.

While you have leadership's attention, get your priority rule signed off, with exploited flaws ahead of everything and internet-facing or identity systems ahead of the CVSS queue, and agree on who can pause and roll back your own changes when a provider is having a bad day. Then measure compliance by what each host needs, counting a server as patched only when every update channel it depends on is covered, and put end-of-support runtimes on the budget agenda, with an owner and an estimated cost for each .NET and PowerShell migration, before an auditor (or an attacker) raises the question first.