A leadership briefing from Complete Network Management (CNM) for our clients.
Keeping It in Perspective
If your spider senses were not going off this week, then you are not paying attention. The goal is to stay alert, not alarmed.
After a wasp buzzes you, you flinch at every breeze or buzzing sound. This week worked the same way. Outages and cyberattacks happen every day. This week, two of them really hurt. Email was down for about a day, and flights were grounded for about four and a half hours. Both were caused by the companies’ own changes, not by attackers. The real attacks this week did not cause any reported operational impact. People online were saying something big was coming, and that put everyone on edge. In a normal week, most of these events would have gone unnoticed.
I work in cybersecurity every day, and the constant stream of alerts and the pace of attacks becomes familiar. For anyone who doesn’t see it daily, this was one crazy week. Even my own instincts were on alert, piqued mostly by the Outlook and FAA outages. The good news is we now have a clearer picture of where our country’s systems are fragile and where they’re not.
If someone were to view the Microsoft and FAA outages as cyberattacks, they had limited effect. They disrupted services, but the impact lasted about a day at most, not days or weeks.
The speed at which we are now finding and patching vulnerabilities with the rise of AI has had a dramatic impact on the IT space. Think about developers needing to write a fix for a zero-day as quickly as possible and get it into production. Then administrators need to apply it just as fast. This shortens the testing window and dramatically increases the likelihood of impacts to production.
An interesting point was brought up, that I find plausible, but we have no facts to back it up, but it’s worth mentioning.
Earlier this year, Project Glasswing was created to give our critical vendors access to find and fix vulnerabilities in their systems through access to advanced AI engines. What if a major vulnerability was detected and a patch had to be rushed out to address it across a broad spectrum of vendors? This would explain why so many systems were “doing maintenance” and patching using “less than optimal change management processes.”
The timeline does line up with the November 17 critical patch release of the Java4Shell vulnerability. So it’s possible, critical vendors got this first and were rushing to patch before the rest of us got it or we might see a new critical vulnerability announced next week. Again, this is speculation and not fact.
It’s our responsibility to learn from this week, stay aware of how the landscape is changing, and adapt our plans accordingly.
— Jason Blanchard, Owner and CEO, Complete Network Management
Summary
- Several notable outages hit from Nov 12 to 17. Every one with a stated cause traced to a change or fault inside the organization’s own systems, not a cyberattack.
- Separately, there were real attacks: a supply-chain attack through Optimizely, an ongoing attack on Equinor, and attacks on ConnectWise tools used by IT providers. Critical Java flaws were also patched.
- CNM has reviewed its exposure to each event; details are in “How CNM Responded” below.
- The lesson: critical services depend on a few systems and vendors, and the pace of change is rising. One faulty update can have national reach.
How We Got Here
This timeline adds the background behind the week of Nov 12–17, oldest first. Several items that made headlines this week are not new. The ScreenConnect flaw was found and patched in September. The ConnectWise Cloud Backup attack happened in October. The Volt Typhoon research and the Pentagon’s base program were public weeks earlier. They were already known to the security community; this week’s outages simply brought them back into the news. Items marked Upcoming have not happened yet.
| Date (2026) | Type | Event |
|---|---|---|
| Sep 8 | Vulnerability | ConnectWise releases ScreenConnect 26.6.5 to fix CVE-2026-84869, a critical client file-transfer flaw; cloud-hosted instances like CNM’s are updated by ConnectWise |
| Sep 11 | Vulnerability | CISA adds the ScreenConnect flaw to its Known Exploited Vulnerabilities catalog |
| Sep 30 | Government | Pentagon announces FORTRESS America to keep military bases running 90 days without the civilian grid |
| Oct 10–17 | Attack | China-aligned Storm-0940 abuses ConnectWise Cloud Backup credentials against 2,250+ Microsoft 365 tenants; ConnectWise shuts the service down Oct 17 (a separate product CNM does not use) |
| Oct 21 | Attack | Microsoft attributes the ConnectWise Cloud Backup attack to Storm-0940 and its use of the ARTEX AI tool |
| Oct 28 | Attack | Lumen’s Black Lotus Labs details Volt Typhoon’s “Silent Train” network and four zero-day campaigns |
| Nov 11 | Attack | ConnectWise reports unauthorized access to some ScreenConnect partner accounts (SCKit) and, as a precaution, disables certain third-party extensions in ~400 partner environments; ConnectWise says having the extensions does not mean an environment was compromised. CNM’s ScreenConnect instance was not among the ~400, and CNM took immediate action to verify and harden its environment. |
| Nov 12 | Outage | Microsoft 365 / Exchange Online outage begins; a separate morning Teams outage hits North America |
| Nov 13 | Outage | FAA nationwide ground stop, about four and a half hours; Microsoft 365 restored |
| Nov 13 | Warning | FBI issues FLASH alert on ARTEX AI attack framework |
| Nov 13–14 | Outage | Workday down from Friday evening into Saturday; many workers could not clock in or handle payroll |
| Nov 14 | Government | White House confirms a Trump–Xi–Putin meeting at the APEC summit in Shenzhen |
| Nov 15 | Government | Homeland Security secretary says Americans do not need to stockpile supplies |
| Nov 15 | Outage | AWS reports authentication, sign-up and connectivity problems in its US East (N. Virginia) region; not tied to the other outages |
| Nov 16 | Outage | JPMorgan Chase software bug delays transactions; customers of other banks report problems through Chase transactions |
| Nov 16 | Outage | United Airlines system outage delays departures; resolved same day |
| Nov 16 | Attack | Optimizely supply-chain attack serves malicious code to visitors of 178 organizations’ websites |
| Nov 16 | Outage | Microsoft publishes preliminary root cause: a latent defect in a July gateway fix |
| Nov 16 | Outage | Anthropic’s Claude Meetings feature could not join new calls due to higher-than-expected demand; other Claude features unaffected |
| Nov 17 | Outage | NSA Net disruption during scheduled maintenance |
| Nov 17 | Outage | Apple services (iCloud, Apple Pay, App Store) down after an internal update; restored by evening |
| Nov 17 | Outage | National Weather Service websites briefly down during routine network maintenance |
| Nov 17 | Outage | Marriott International reported among the day’s outages |
| Nov 17 | Attack | Equinor (Norwegian energy company) detects an ongoing cyberattack on some IT systems; production unaffected |
| Nov 17 | Vulnerability | Java4Shell disclosed; Oracle and OpenJDK ship fixes in the November patch update |
| Nov 17 | Government | FEMA and FCC nationwide Emergency Alert System test; the date social media rumors claimed a 90-day blackout would begin |
| Nov 19 | Upcoming | Trump–Xi–Putin meeting in Shenzhen |
| Nov 20 | Upcoming | Microsoft’s final post-incident report due |
The Outages
| Event | Date | Cause | Cyberattack? |
|---|---|---|---|
| Microsoft 365 | Nov 12–13 | A July fix to gateway servers failed under heavy load; 80%+ of those servers became unresponsive | No |
| FAA ground stop | Nov 13 | New L3Harris router in Arkansas flooded the network; 17 of 22 en route centers hit. Final report pending | No |
| JPMorgan Chase | Nov 16 | Software bug in on-premise hardware. Other banks were affected only through Chase transactions. Resolved that evening | No |
| United Airlines | Nov 16 | Widespread system outage delayed departures; resolved same day; cause not stated | No indication |
| NSA Net | Nov 17 | Scheduled routine maintenance (NSA statement); sources say patching | No evidence |
| Apple | Nov 17 | Internal update; services restored | No |
| National Weather Service | Nov 17 | Routine network maintenance | No |
On Microsoft, IT professionals point to 2026 buyouts of senior engineers and lost institutional knowledge; Microsoft has not linked staffing to the outage. Uptime Institute data shows configuration and change-management failures are the most common cause of major network-related outages (50%).
The Attacks and Vulnerabilities
| Event | Date | What happened |
|---|---|---|
| Optimizely supply-chain attack | Nov 16 | Malicious code in a website testing tool hit visitors of 178 organizations (HSBC, Equifax, CNN, Zoom and others); could steal card data and swap crypto wallet addresses |
| Equinor | Nov 17 | Ongoing attack on Europe’s largest gas supplier; no data theft or production impact so far; attacker unknown |
| ConnectWise | Oct–Nov | China-aligned attackers targeted 2,250+ Microsoft 365 tenants using stolen ConnectWise Cloud Backup credentials (Oct). Separately, attackers hijacked ScreenConnect partner accounts to push rogue remote-access software; ConnectWise disabled suspect extensions in about 400 partner environments (presence alone does not mean compromise) |
| FBI warning: ARTEX | Nov 13 | AI attack tool used by criminals and state groups since August, including against financial services; some intrusions reached admin access in under three hours |
| Volt Typhoon (China state-sponsored group) | Reported Oct 28 | A Chinese state-sponsored group focused on U.S. critical infrastructure that quietly gains and keeps access rather than causing visible damage. Lumen’s Black Lotus Labs found it hid its activity behind a rotating network of hijacked home and business devices it calls “Silent Train” and used it in four zero-day campaigns against tools that manage many networks at once: Versa Director, N-able N-central, Cisco Catalyst SD-WAN and VeloCloud Orchestrator. All four vendors have released fixes |
| Java4Shell and Oracle patches | Nov 17 | Critical Java flaw (9.2 per Calif; 8.1 per Oracle), found with AI help; affects Java 17, 21, 25 and 27 plus Java 8 Performance Pack; patched Nov 17; Calif says there is no universal exploit |
How CNM Responded
Here is where each relevant event touches our stack.
| Event | Does it affect CNM? | What we are doing |
|---|---|---|
| ConnectWise Cloud Backup breach (Storm-0940) | No. CNM does not use ConnectWise Cloud Backup (formerly SkyKick) | No action needed |
| ScreenConnect partner-account attacks (SCKit) | Partly. Our instance is cloud-hosted, so ConnectWise patches the server as soon as fixes are available. | Confirming every technician client is on version 26.6.5 or later, checking for the flagged extensions, enforcing MFA on every account, systems are behind allowlists and ZTNA, reviewing session logs, and hunting client endpoints for any ScreenConnect agent that does not belong to CNM and removing any we find. |
| Java4Shell | Depends on each client’s software | Identifying internet-facing systems that run affected Java versions and confirming vendor patches |
| ARTEX (FBI warning) | Applies to everyone; attacks can reach admin access within hours | Using Field Effect Complete, Huntress and SentinelOne to hunt for the FBI’s published indicators, and prioritizing leaked credentials and exposed remote-access tools |
The 90-Day Blackout Rumor
- Social media claims a 90-day nationwide blackout began Nov 17. There is no evidence; a video of the President announcing it was faked.
- What’s real: a routine FEMA alert test on Nov 17, and a Pentagon program (FORTRESS America) to keep military bases running if the grid fails.
- The Homeland Security secretary said Americans do not need to stockpile.
Why Patching Is Speeding Up
AI tools are now finding software flaws faster than ever. The number of publicly disclosed vulnerabilities in 2026 passed 100,000 on Nov 13, more than double all of 2025. Microsoft’s June update fixed 206 flaws in a single month, a record. This week’s Java4Shell flaw was found with AI help. Not every new flaw is AI-found, and no one tracks exactly how many are, but the direction is clear. The UK’s National Cyber Security Centre has told organizations to prepare for a “patch wave,” and CISA now gives federal agencies as little as three days to fix some actively exploited flaws.
The Bigger Picture: The Power Grid
Everything in this document runs on electricity. If the power grid fails, the internet, cloud services and phone networks go down with it. Nothing this week involved the power grid. But the risk is real enough that the Pentagon launched FORTRESS America to keep military bases running for 90 days without the civilian grid, and Volt Typhoon’s campaigns show foreign groups working to position themselves inside U.S. critical infrastructure.
In south Louisiana, we already know how to handle this. Every few years a hurricane takes out power, internet and phone service for days at a time, and we plan and practice for it. The same preparation that gets us through hurricane season protects against a grid failure from any cause:
- Backup power for critical systems, and fuel to keep it running
- More than one way to reach the internet, such as a second provider or a cellular or satellite backup
- Offline and offsite copies of critical data
- A tested plan for how your team works, communicates and serves customers without the cloud for a day or more
- Contact lists and key procedures on paper
A national grid failure would be harder, but we are certainly better prepared than most.
There are good reasons to live in Louisiana, and the people are the main reason. We get to practice being a real community during hurricanes. We are reminded of the benefits of working together and helping one another.
Common Threads
- Concentration: a few providers carry much of the economy; one failure has national reach.
- Change risk: most outages followed updates, patches or new equipment. More patching is coming as AI finds more flaws.
- Trusted vendors as the way in: Optimizely and ConnectWise show attackers going through suppliers.
- Speed: AI tools like ARTEX shorten the time from first contact to full access to hours.
Closing
- A week of high-profile outages, mostly self-inflicted by system changes, alongside real attacks through trusted vendors.
- None of this week’s events calls for panic. It does call for steady attention to vendors, change control and fast detection, which is the work CNM does for you every day.
- Questions about how any of this affects your organization? Contact CNM and we’ll walk through it with you.
Sources
Outages
- Microsoft 365: Heise – Latent bug in July fix took down gateways; The Stack – Latent bug knocked 80% of networking servers offline; AP via WSVN – Outage hits Outlook
- Microsoft staffing: Computerworld – Microsoft buyouts for ~7% of U.S. workforce; Fast Company – Outage puts layoffs under scrutiny; Network World – Uptime Institute outage report
- FAA ground stop: KATV – El Dorado, Arkansas at center of breakdown; Business Traveller – What caused the FAA shutdown; Bernama – Investigation status
- JPMorgan Chase: Yahoo Finance – JPMorgan says issues resolved; Tech Times – Bug slowed rival banks only through Chase transactions
- United Airlines: CNBC – United resolves system outage
- NSA Net: The Record – NSA Net disruptions; CBS via KTSA – NSA network goes down
- Apple: Reuters via U.S. News – Outage caused by internal update; AOL/CBS – Apple restores services
- National Weather Service: NWS on X
Attacks and vulnerabilities
- Optimizely: SecurityWeek – Optimizely vulnerability exploited; Reflectiz – Crypto swapping and card skimming
- Equinor: Equinor – Company statement; TokenPost – Europe’s largest gas supplier
- ConnectWise Cloud Backup: Dark Reading – Microsoft links Storm-0940 to ConnectWise attack
- ConnectWise ScreenConnect: BleepingComputer – Active ScreenConnect attacks; The Register – ConnectWise disables extensions in 400 environments; ConnectWise – ScreenConnect 26.6.5 security bulletin
- ARTEX: BleepingComputer – FBI warns about ARTEX; OODA Loop – FBI FLASH-20261113-001
- Java4Shell: SentinelOne – Java4Shell; CyberScoop – Found with help of AI; Cycode – Affected versions; SecurityOnline – No universal exploit
- Volt Typhoon management-system campaigns: Help Net Security – Volt Typhoon Silent Train
Blackout rumor
- Fact check and DHS stockpile comments: PBS NewsHour
- FEMA alert test: FCC Public Notice DA 26-989
- Pentagon base program: The War Zone – FORTRESS America
Timeline (additional)
- ScreenConnect CVE-2026-84869 and CISA listing: Gridinsoft – ScreenConnect enters KEV
- ConnectWise Cloud Backup dates and attribution: SecNews – Storm-0940 and ARTEX; Neoteo – Microsoft ties Storm-0940 to sign-in attempts
- ScreenConnect SCKit disclosure: Field Effect – SCKit campaign
- Trump–Xi–Putin meeting confirmed: AP via CP24
- Marriott and other Nov 17 outages: Fox News – Tech outages live updates
- Microsoft final report due Nov 20: Heise – Exchange Online outage
Why patching is speeding up
- The Register – This year’s bug count just passed 100K
- Cloud Security Alliance – AI-accelerated vulnerability discovery and patch debt
- Safeguard – How much growth is AI-found
- Security Affairs – UK NCSC warns of a patch wave
- Executive Gov – CISA three-day deadlines
The power grid
- The War Zone – FORTRESS America
- CSO Online – Volt Typhoon pre-positioning in critical infrastructure
Workday, AWS and Anthropic outages