Microsoft, FAA and US Bank Outages in Five Days Put DDoS Resilience in Focus
Key Takeaways
- •The FAA's failure description, involving a newly installed router firing roughly 1,000 connection attempts every four to five seconds until it overwhelmed network equipment, matches the textbook definition of a denial-of-service attack, according to MazeBolt CEO Matthew Andriani.
- •None of the three organizations had disclosed a confirmed root cause, with Microsoft citing only a network upgrade, Chase describing an undisclosed technical issue, and the FBI still involved without ruling out a cyberattack in the FAA case.
- •A MazeBolt-commissioned survey of 300 chief information security officers at banks, insurers and payment providers found that 86% test their DDoS defenses once a year or less and only 5% are fully confident those defenses would stop an attack.
- •In MazeBolt's first validations of enterprise production environments, an average of 37% of attack vectors got through protections that were already in place and believed sound.
- •Andriani advised boards to stop asking whether they own DDoS protection and instead demand evidence of resilience certainty gathered continuously in production across the entire attack surface.

In the span of five days in November 2026, three pillars of American life stopped working. On Thursday, Nov. 12, Microsoft Outlook went down worldwide, generating more than 43,000 problem reports from 72 countries by the following morning in an outage that stretched past 12 hours. On Friday, the Federal Aviation Administration (FAA) halted every departure in the United States for more than four hours after its communications network failed; nearly 12,000 flights were canceled, and the backlog was still being felt days later. On Monday, customers of Chase, Bank of America, Wells Fargo, U.S. Bank and Truist found themselves locked out of online banking, with Chase alone drawing more than 5,000 outage reports by early afternoon.
Each organization was quick to say what the disruptions were not: attacks.
Microsoft stated, “There is no indication this is the result of malicious or cyber activity.” Chase said, “This is a technical issue, not cyber related.” Homeland Security Secretary Markwayne Mullin said there was no indication of a “nefarious attack” on the FAA.
But a closer look at what each organization described tells a different story, according to Matthew Andriani, chief executive of DDoS resilience firm MazeBolt, writing in a Fortune commentary. Transportation Secretary Sean Duffy said, “Our switches, our routers, actually got overloaded with data,” after a newly installed router fired roughly 1,000 connection attempts every four to five seconds. A flood of requests overwhelming network equipment until legitimate traffic cannot get through is not merely similar to a denial-of-service attack, Andriani argued — it is the textbook definition of one.
FAA Administrator Bryan Bedford has said that explanation does not fully account for the root cause. Microsoft has blamed a network upgrade but has not published a full post-incident report. Chase has not disclosed what its “technical issue” was. Those pending disclosures are what to watch next.
Services unreachable. Systems saturated. Recovery measured in hours. No root cause disclosed. In each case, the fast answer was the same: no cyberattack. As Andriani put it, in reality it looks like a duck, walks like a duck and quacks like a duck.
Then there is the timing. Any one of these failures on its own could be bad luck. But three outages across cloud software, aviation and banking in five days form a pattern — and in any other security context, a pattern triggers an investigation, not a press release, he argued. “No indication” on day one is not the same as “ruled out” after forensics. On the day of the FAA failure, officials initially declined to rule out a cyberattack. That was the right instinct. Days later, with the FBI involved, a cyberattack has still not been ruled out.
There is a reason the word DDoS rarely appears in early statements, Andriani wrote. A “technical issue” is a maintenance problem. An “attack” is a disclosure event, one that prompts regulators, investors and customers to ask how it got through. Every incentive points to the first label, whether or not anyone acts on it.
DDoS — distributed denial of service — is a leading threat to the availability of any publicly accessible online service. Every bank, airline and cloud provider runs hundreds or thousands of services, and each is a potential target. For everyone else, Andriani noted, the cause matters less than the impact: an hour offline costs the same whether a router misfired or a botnet fired. A DDoS attack is the cheapest, most repeatable way for an adversary to buy that hour. For a bank, it means frozen payments, regulatory notifications and customers wondering where else to keep their money.
Every executive should be asking one question this week, he argued: if someone tried, would our systems hold? Most cannot answer it.
Andriani’s company, MazeBolt, provides DDoS resilience certainty to its customers. It commissioned a survey of 300 chief information security officers at banks, insurers and payment providers, which found that 86% test their DDoS defenses once a year or less, and only 5% are fully confident those defenses would stop an attack. In his experience, reality tends to shock those who believe a few tests a year are enough: in MazeBolt’s first validations of enterprise production environments, an average of 37% of attack vectors got through protections already in place and believed sound. An enterprise that cannot vouch for what it has bought, he wrote, is carrying risk it has never measured.
No one in the value chain will say so, he added. The mitigation vendor reports what its system sees, not what it misses. The cloud provider’s contract covers its infrastructure, not the customer’s configuration. The auditor confirms a control exists, not that it works on the day it is needed. Each answers its own question honestly; none answers the customer’s.
The warning comes in the era of AI-orchestrated attacks. Andriani said the length of DDoS attacks is increasing and that additional shifts on the attackers’ side are certain to follow. To deal with the new reality, organizations will need to proactively map out their vulnerabilities and strive toward proper resilience.
His advice to boards: stop asking whether they own DDoS protection, because nearly every large enterprise does. They should instead demand evidence of DDoS resilience certainty, gathered continuously in production, across the entire attack surface. That certainty, he argued, is what will keep payments flowing, planes moving and customers logged in. Organizations that hold it will spend the next bad week reading the headlines instead of appearing in them.
The opinions expressed in Fortune.com commentary pieces are solely the views of their authors and do not necessarily reflect the opinions and beliefs of Fortune.
This story was originally featured on Fortune.com.