FAA Ground Stop, Microsoft 365, Chase and NSA Outages: What We Know About Four Failures in Six Days

Chris Rees
25+ years in IT · Pluralsight author, 4.6/5 across 2,000+ ratings

Between November 12 and 17, 2026, four critical systems failed: Microsoft 365 on Thursday, November 12; the FAA's air traffic control telecom network on Friday the 13th; Chase's online banking on Monday the 16th; and the NSA's classified internal network on Tuesday the 17th. It looks like a coordinated attack, and plenty of people online have said so. The evidence so far says otherwise. Each outage has its own, mostly mundane, cause, and at least three of the four trace back to a change someone made to a working system. This post goes through each one: what happened, the technical cause as currently stated, what is still disputed, and what it teaches anyone who runs, defends or studies production systems.
Microsoft 365 outage (November 12–13): a latent bug meets a bad day
What happened. At 16:20 UTC on Thursday, November 12 (11:20 a.m. ET), Microsoft opened incident EX1550543 for Exchange Online, and two hours later MO1550821 for Teams, Copilot and Defender. Many users could not reach their mailboxes, and mail delivery stalled. Exchange took the main hit; the list of affected services grew to include SharePoint, OneDrive, Purview, Intune, Planner, Power BI, Fabric and more, though Microsoft says Teams chats and meetings worked for most of the outage. It began in the Americas and spread to every region; Downdetector logged reports from 74 countries, and US reports for Outlook alone peaked around 130,000 in the early afternoon Eastern time. Azure was not among the affected services. Microsoft declared the incident over at 16:10 UTC on November 13, just under 24 hours after it started, though backlogged mail and final updates ran into Saturday.
The technical cause. Microsoft's preliminary post-incident review (PIR), dated November 15, describes three factors that together took down "the networking gateway that fronts many Microsoft 365 services":
- A latent defect from July. A change to the gateway servers, deployed starting in July, introduced lock contention inside an internal Windows process under certain load conditions. The July change was itself a fix built only for the M365 gateway servers; Microsoft says neither the defect nor that fix applies to broader Windows deployments. It passed testing because the contention is worse on newer machines with more cores, and the load tests ran on weaker hardware with fewer cores and a different traffic mix. It then ran worldwide for months without trouble.
- Extra load. At 01:07 UTC on November 12, about 15 hours before the outage, an automated failover moved Teams traffic onto gateway servers shared with Exchange Online.
- Reduced capacity. Meanwhile, nearly 14% of the gateway's server capacity was out of service because of an unrelated issue from the day before. On its own, that was within the service's headroom.
Together, the extra load and the missing capacity pushed the servers into the conditions that triggered the July bug, during peak US traffic. Requests piled up until servers stopped responding. Then the system's own resilience made it worse: automatic rerouting sent each failed server's traffic to the healthy ones, which pushed them over too. In under two hours the failure had spread across the fleet, and at the worst point more than 80% of the gateway servers were unresponsive. The bulkheads that normally keep a failure in one region from spreading did not hold, because of a separate issue. And because the defect had shipped months earlier, Microsoft says its normal rollback, flighting (staged release rings) and feature flags could not quickly revert it. Servers restored around 23:55 UTC on November 12 stayed healthy at first, then became unresponsive again once traffic reached them. The eventual fix, a code revert first applied at 07:20 UTC on November 13, returned the code path to its previous known-good state. Europe was healthy by 11:35 UTC and webmail was normal worldwide by 14:17 UTC.
What is disputed. Microsoft's explanation changed. On November 13 it pointed to "an issue during an earlier network capacity upgrade," and early reports described rolling back a recent network configuration change. The PIR describes a code revert of a July Windows change instead. The PIR is preliminary; Microsoft says the final version, with a full list of repair items, is due November 20. Microsoft said it found no indication of malicious or cyber activity.
What it teaches. This is the textbook shape of a large outage: no single cause, but a latent defect, a load shift and a capacity shortfall lining up at once, then a failover mechanism spreading the damage. Two details are worth remembering. First, a test lab that doesn't match production hides bugs: this one was worse on newer, many-core hardware than on the machines used for load testing. Second, rollback tooling is built for recent changes. A defect that sits quietly for four months has aged out of it, so you need a way to revert old changes too.
FAA ground stop (November 13): what caused it
What happened. Shortly after 3:30 p.m. ET on Friday, November 13, radar and radio data feeding US air traffic control became unstable. At 3:54 p.m. ET the FAA ordered a nationwide ground stop, which it lifted at 8:15 p.m. ET, about four hours and 20 minutes later. The instability hit 17 of the 22 Air Route Traffic Control Centers (ARTCCs) plus several approach control facilities. The FAA stresses that this was not a full outage: some frequencies and radar feeds remained reliable. About 3,500 aircraft were airborne when it started, and all landed safely. The disruption rolled through the weekend: more than 9,000 cancellations on Friday (the FAA notes that not all were directly caused by the ground stop), about 6,500 on Saturday and more than 2,000 on Sunday, according to FlightAware data reported by CNBC. FedScoop puts the three-day total near 18,000; all of these figures are approximate.
The technical cause. The FAA's voice and data circuits run over the FAA Telecommunications Infrastructure (FTI), a network owned and operated by contractor L3Harris. L3Harris is converting FTI from copper to fiber and installing new Juniper routers as it goes. The FAA has identified two potential sources of the trouble: a radar site in Amarillo, Texas, whose router was replaced about three weeks earlier, and a radio communications site in El Dorado, Arkansas, with a newly installed router. According to the FAA, the routers flooded the network with traffic at a rate of 1,000 connection attempts every four to five seconds, enough to destabilize the circuits carrying radar and radio data to the centers. The network stabilized after L3Harris blocked the two sites.
The FAA has since disabled a broadcast function on more than 60 other Juniper routers and has paused new Juniper installations (reported November 17); sites stay on their older Cisco routers in the meantime. The scale of the rollout puts that in context: Senate Commerce Committee Democrats say 3,648 routers are due to be replaced, about 400 have been installed and only 69 are operational.
What is disputed. Almost everything about why. Transportation Secretary Sean Duffy initially blamed aging air traffic control technology and overloaded routers; Senate Democrats point out that the routers implicated are the new ones. Duffy has since asked whether it was an installation error by L3Harris or a faulty router, while the FAA's own Q&A attributes the instability to the legacy TDM system rather than the upgrades. Hewlett Packard Enterprise, which owns Juniper, says its device did not fail or act improperly and functioned as it was configured to by the installer. Professional Aviation Safety Specialists (PASS), the union for FAA technicians, says its members were not involved in the upgrades. In a November 16 letter, 13 Senate Commerce Democrats led by Sen. Maria Cantwell asked for a pause on installs, a technical briefing and written answers by November 23, noting that identifying the device that failed does not establish why it failed.
On cyberattack, the FAA says there is no indication of one or of an outside threat, and Bedford said there was "zero evidence" of anything nefarious. The routers are being tested at L3Harris's lab in Florida with FAA staff and FBI agents present; the FBI says it is working with DOT, the FAA and L3Harris but has not taken possession of any equipment.
A separate, brief United Airlines IT outage on November 16 delayed departures. CNBC reported no sign that it was related to the ground stop.
What it teaches. A device flooding a network with connection attempts or broadcasts is a classic failure mode, and networks have controls for it: storm control, control-plane rate limiting, and segmentation that keeps one site's noise from reaching every center. The bigger lesson is about migrations. During a copper-to-fiber transition, the network is a hybrid of old and new, and new equipment talking to old equipment is exactly where untested behavior lives. HPE's point that the router did what it was configured to do is the reminder: configuration is part of the change. Staged rollouts with monitoring, and the ability to isolate a site quickly, are what turn a bad router into a local problem instead of a national one.
Chase outage (November 16): one bug, symptoms at many banks
What happened. On Monday, November 16, Chase customers began reporting problems with logins, balances, transfers, withdrawals and Zelle payments. Downdetector reports for Chase climbed from shortly after 9 a.m. Eastern time; outlets cite figures from about 4,500 to nearly 5,000 around 1:30 to 2 p.m. ET, and Fortune reported a much higher running count by 2:22 p.m. Reports also spiked for Bank of America (around 2,000), U.S. Bank (around 1,000) and Wells Fargo (just under 1,000). Chase said customers still had access to their accounts and funds, though some transactions were taking longer than usual. By about 6:30 p.m. ET a Chase spokesperson said transactions were posting normally.
The technical cause. Chase attributed the problem to "a temporary issue related to a software bug in on-premise hardware system." It said the problem was "not cyber-related" and "not a systemic issue." Chase has not named the system or the vendor. An unnamed FBI official told Fox Business the disruption traced back to a software release the bank rolled out earlier that Monday and that the impact was limited to Chase; Chase has not confirmed that account.
Why other banks showed up. Chase said other banks were affected only to the extent they were a party to a transaction with Chase. U.S. Bank said it was not directly impacted and that delays involved Zelle payments with other banks. PNC, Truist and Capital One said they had no direct impact, Bank of America later said its systems operated normally throughout the day, and Zelle said its own service was operating normally. In other words, the spikes at other banks look like a downstream effect: if your payment to a Chase customer stalls, you report your own bank as down.
What is disputed. Downdetector counts and start times vary by source. No regulator has commented on the record. The software-release account comes from an anonymous official and has not been confirmed by Chase. Claims that the outage was linked to the FAA and Microsoft incidents are unconfirmed.
What it teaches. Crowd-sourced outage data measures symptoms, not causes. A failure at the largest participant in a payment network will show up as "outages" at every bank that transacts with it. Incident responders need to separate where users feel the pain from where the fault is, and the same discipline applies to reading Downdetector during your own incident.
NSA network outage (November 17): NSANet degraded during maintenance
What happened. On Tuesday, November 17, CNN, CBS News, Nextgov and others reported that NSANet, the NSA's classified internal network and its primary platform for top-secret communications, was disrupted. Some staff could not reach their workplace systems, and according to CBS sources the agency staggered arrival times for non-emergency staff at Fort Meade so that everyone would not log on at once. The latest reporting says restoration was ongoing but not complete.
The stated cause. An NSA spokesperson said the agency experienced temporary degradation during scheduled routine maintenance, that some internal systems were affected, and that mission-critical capabilities remain fully operational. Everything beyond that comes from anonymous sources: two told CBS it was routine system patching, four of six sources told Nextgov it may be linked to a series of software upgrades, and one told the Wall Street Journal the agency was updating software to patch the system.
What is disputed. How it happened and how long it will last. No outlet has said exactly when the network went down or which services were lost. On cyberattack, the wording is "no evidence," not "ruled out": a White House spokesperson said the administration sees no evidence it was a cyberattack, and the NSA's own statement does not mention one. No official has linked it to the other outages.
What it teaches. The NSA is subject to the same rule as everyone else: maintenance is a change, and changes break things. Classified networks are usually air-gapped from the internet, which limits outside threats but also limits options during recovery, such as falling back to a cloud service or pulling a fix from a vendor in minutes. The staggered arrivals are a practical recovery technique worth noting: when a system comes back, a thundering herd of simultaneous logins can knock it over again, just as returning traffic did to Microsoft's restored servers.
Were the outages connected, or a coordinated cyberattack?
Nothing published so far supports that, and Homeland Security Secretary Markwayne Mullin called the incidents individually unrelated. Each organization has given a non-malicious cause, and the causes are different in kind: a Windows code defect at Microsoft, router behavior on an FAA contractor's network, a software bug in Chase's on-premises hardware and scheduled maintenance at the NSA. They also span unrelated infrastructure: a global SaaS platform, a contractor-owned FAA telecom network run by L3Harris, a bank's own data centers and an air-gapped classified network. A single attacker would need a foothold in all four.
Two caveats. First, "no evidence of a cyberattack" is not the same as "ruled out," and the FAA's root cause in particular is still under investigation, with the FBI involved. Second, early explanations have already shifted (Microsoft's did), so the final reports matter. But the pattern here is a familiar one: large systems fail all the time, and a cluster of failures in one week is usually coincidence plus heightened attention, not a campaign.
The common thread: change management and blast radius
| Outage | Trigger | Why it spread | Recovery |
|---|---|---|---|
| Microsoft 365 | Latent July code change, plus failover load and 14% capacity loss | Automatic rerouting cascaded the failure; regional bulkheads failed; rollback could not revert a months-old change | Code revert, staged bring-up |
| FAA | New routers on the FAA telecom network flooding it with connection attempts | One network carries radar and radio data for most centers | Blocked two sites; disabled broadcast on 60+ routers; paused installs |
| Chase | Software bug in on-premises hardware | Payment counterparties (Zelle, transfers) felt it | Transactions posting normally by that evening |
| NSA | Scheduled maintenance (patching or upgrades per anonymous sources) | Network-wide platform | Staggered logins, ongoing restoration |
Strip away the headlines and three lessons are left:
- Every change is a risk, including the routine ones. Patching, upgrades and hardware swaps caused or set up most of these outages. Change management is not paperwork: it is impact analysis, a tested backout plan and a maintenance window sized to the risk.
- Contain the blast radius. Bulkheads, segmentation, storm control and rate limiting exist so one bad component becomes a local incident. When they fail, so does everything behind them, and failover can carry the fault to the next healthy server.
- Plan for the restart, not just the failure. Returning traffic knocked Microsoft's restored servers over again; the NSA staggered logins to avoid the same thing.
What this means for your certification
Security+. Availability is one-third of the CIA triad, and this week was an availability case study. SY0-701 objective 1.3 covers change management (approval, impact analysis, backout plans, maintenance windows, testing) and 3.4 covers resilience (high availability, capacity planning, testing and recovery). Our posts on change management and resilience and high availability map directly to what failed here.
CySA+. The skill on display is triage: separating the fault domain from the symptoms, and deciding quickly whether an event is malicious. Review the incident response lifecycle and declaration, escalation and lessons learned; Microsoft's PIR is a good model of a lessons-learned report.
CISSP and CCSP. Business continuity, disaster recovery and cloud resilience. Microsoft's outage is a shared-responsibility question for every customer (what is your plan when your SaaS provider is down for a day?), and the FAA's is a third-party and supply-chain risk question. See disaster recovery testing and cloud resilience and DR patterns.
FAQ
Were the FAA, Microsoft, Chase and NSA outages connected?
No connection has been found. Each organization gave a different, non-malicious cause: a latent Windows code defect at Microsoft, newly installed routers on the FAA's contractor-run telecom network, a software bug in Chase's on-premises hardware, and scheduled maintenance at the NSA. No official has linked any of the four, and none has been attributed to a cyberattack.
What caused the FAA ground stop on November 13, 2026?
The FAA has identified two potential sources: newly installed Juniper routers at a radar site in Amarillo, Texas, and a radio site in El Dorado, Arkansas, on the L3Harris-run FAA telecom network. The FAA says they flooded the network with about 1,000 connection attempts every four to five seconds, destabilizing radar and radio data to 17 of 22 en route centers. Why the routers misbehaved is still under investigation.
Was the FAA ground stop a cyberattack?
The FAA says there is no indication of a cyberattack or outside threat, and FAA Administrator Bryan Bedford said there is zero evidence of anything nefarious. The root cause is still under investigation, the FBI is working with the FAA and L3Harris, and senators have asked for written answers by November 23.
What caused the Microsoft 365 outage on November 12, 2026?
On November 12–13, 2026, according to Microsoft's preliminary post-incident review, a latent lock-contention defect in a Windows process on its M365 gateway servers, introduced in July, was triggered when a Teams failover added load while about 14% of gateway capacity was already offline. Automatic rerouting spread the failure until more than 80% of gateway servers were unresponsive. Microsoft fixed it by reverting the code path.
What caused the Chase outage on November 16?
Chase blamed a temporary issue related to a software bug in an on-premise hardware system, said it was not cyber-related, and has not named the system or vendor. Other banks saw reports mainly because their customers' payments to and from Chase stalled.
Did Chase customers lose access to their money?
Chase said customers retained access to their accounts and funds, but many reported failed logins, delayed transfers and Zelle problems for several hours on November 16. Chase said transactions were posting normally by that evening.
Is the NSA's network back up?
The NSA says NSANet was temporarily degraded during scheduled routine maintenance; anonymous sources describe patching or software upgrades. As of November 17, restoration was ongoing but not complete. The NSA says mission-critical capabilities remain fully operational.
- Audit your last ten changes for a real backout plan. For each one, ask whether you could revert it in under an hour, including changes that went live months ago. If the answer is no, fix the plan before the next change.
- Write down your plan for a 24-hour email outage. Exchange Online was down for many Microsoft 365 customers for most of a day. Decide now how your team communicates, who declares the incident and where the out-of-band contact list lives.
- Check that one noisy device cannot take down a site. Confirm storm control and control-plane rate limiting are enabled on your switches and routers, and that a new device is staged and monitored before it joins production.
Sources: Microsoft preliminary post-incident review for EX1550543/MO1550821 (November 15, 2026); FAA statements as reported by PBS NewsHour, CNBC, ABC News and FedScoop; HPE and PASS statements as reported by PBS and others; Senate Commerce Committee Democrats' letter to DOT and FAA (November 16, 2026); Chase, U.S. Bank and Zelle statements as reported by CNBC, NBC News and Fortune; NSA and White House statements as reported by CNN, CBS News, Nextgov/FCW and the Wall Street Journal.
#FAA #Microsoft365 #Outage #Chase #NSA #ChangeManagement #Resilience #IncidentResponse #SecurityPlus #CySA #CISSP #CyberSecurity
Turn this into exam points
Every guide walks the published objectives in order, with exam-style questions you answer in the browser and a PDF to keep. Security+ Domain 1 is free, in full.
- Find your gapsA free readiness quiz scores you domain by domain.
- Study what's on the testWork through every objective, with practice questions.
- Book it with confidenceSchedule the exam when you're scoring well in every domain.
Keep reading
Appetite, Tolerance, Treatment, and the Clause You Forgot: Cloud Risk Management and Contract Design for CCSP
In the cloud, contract is what you have instead of control, and a term you did not negotiate is a control you do not have. CCSP objectives 6.4 and 6.5 from the enterprise risk seat: assessing a provider's own risk program, appetite versus tolerance versus threshold, owner and controller risk against custodian and processor risk, the four treatments as cloud reshapes them, ISO 31000, NIST SP 800-37 and ENISA, key risk indicators, and the contract clauses grouped by what each one protects.
Read CISSP deep divesSynthetic Transactions, Misuse Cases, and Coverage Analysis: The Ten Security Controls Tests CISSP Makes You Tell Apart
CISSP objective 6.2 names ten testing techniques, and the exam rarely asks what one is; it describes a situation and asks which one fits. This post sorts the list by what each technique proves: log reviews and synthetic transactions for controls in operation, code review from static analysis to fuzzing and pair programming, misuse cases, coverage analysis, interface testing, breach attack simulation, and compliance checks, plus the method for choosing the test from the stem.
Read AI attacks & defensesFBI FLASH on ARTEX: How Attackers Turned an AI Pentesting Framework Into an Intrusion Engine, and How to Defend
The FBI's November 13, 2026 FLASH warns that ARTEX, an open-source agentic AI penetration testing framework, is being used to go from first scan to privileged access in under three hours. How ARTEX works, the TTPs mapped to MITRE ATT&CK, the IOCs, detection queries, the FBI's mitigations and three things defenders can do today.
ReadFollow Skillthropic
Keep learning between study sessions
The same exam-focused teaching, in smaller pieces, wherever you scroll.
Enjoyed this? Get the AI security news that matters.
Join The AI Security Brief for the top AI security news, plus what's important to the C-suite. Free, straight to your inbox.
No spam. Unsubscribe anytime.