Tuesday, November 17, 2026

They Came One After Another

The timing stood out before any connection did.

Microsoft 365 went down on November 12. Outlook, Teams, and Copilot stopped working for a large share of users. The company later said a latent defect in its networking gateway servers, introduced in a July update, combined with a Teams failover and reduced capacity to knock more than 80 percent of those servers unresponsive. Microsoft said the outage was not related to a cyberattack or AI. Services were largely restored by Saturday. The Guardian and Al Jazeera covered the initial disruption and Microsoft’s early statements.

The next day the FAA ordered a nationwide ground stop that lasted more than four hours. Transportation Secretary Sean Duffy said contractor L3Harris was replacing old copper lines with fiber at a remote radio site in El Dorado, Arkansas. A newly installed router began flooding the network with more than 1,000 connection attempts every four to five seconds. Radar feeds and communications became unstable at most of the air route traffic control centers. The FAA grounded departures as a safety step. FlightAware recorded thousands of cancellations that day, with more disruption into the weekend. Officials said there was no evidence of a cyberattack. Aviation Week, BBC, and the FAA’s own Q&A laid out the details.

On November 16, Chase customers reported trouble logging in, moving money, and completing transactions. Downdetector showed a sharp spike. Chase said the cause was a software bug in on-premise hardware systems and that it was not cyber-related. The bank described the problem as temporary and non-systemic. Other banks and Zelle users mainly saw delays when a transaction involved Chase. Customers still had access to their accounts and funds, according to the bank. It later said the issue was resolved. Yahoo Finance, Politico, and Reuters reported the statements and the user reports.

The same afternoon United Airlines had a shorter system outage that delayed mainline departures. Pilots said they could not get the final weight-and-balance numbers they need before takeoff. United said the problem was not cybersecurity-related and that flights already in the air and United Express flights were unaffected. The airline was still recovering from Friday’s ground stop. CNN and Reuters carried the airline’s account.

Workday also had a disruption around November 13 tied to its disaster-recovery systems, though that one drew less attention.

This morning the pattern continued. Multiple sources told CNN that the National Security Agency’s highly classified internal communications network, NSA Net, is down or intermittent. The network is what NSA employees use to send classified information to one another. The agency has staggered arrivals at Fort Meade while it assesses the scope. Sources said there is no evidence of a cyberattack at this point; some indicated the problem may be tied to software upgrades. Critical external channels and the ability to reach the president while he is traveling remain intact through redundancies, according to the same accounts. The NSA has not issued a full public statement. CNN, Reuters, New York Post, Nextgov, and The Record have the reporting.

Here's CNN reporting this story: [VIDEO]

No company or agency has linked its failure to the others. The Microsoft problem was a dormant bug plus load conditions. The FAA problem was a routine copper-to-fiber upgrade that flooded the network. Chase pointed to a software bug in its own hardware systems. United’s issue sat in its internal tools for weight data and crew communications. The NSA disruption is still being assessed and has not been tied to the earlier ones. The clustering is what made the week feel like one event.

USA Today and The Independent noted the public reaction and the expert comments that followed. The common thread they describe is concentration and complexity: a single latent defect, a single misbehaving router, or a single internal bug can reach a large number of people quickly when the systems are tightly coupled. That is a structural observation, not a claim that the incidents shared a cause.

The week showed how visible those failures become when they arrive close together. It did not show that they were the same failure.

No comments:

Post a Comment

Comments are now moderated because one random commenter chose to get comment happy. What doesn't get published is up to my discretion. Of course moderating policy is subject to change. Thanks!