Every bankruptcy case runs on notices most people never think about: the notice that tells a creditor a case exists, tells a debtor when and where to appear, and starts the clocks that end rights. Since October 9, those notices have not moved. The Bankruptcy Noticing Center — the one operation that produces and transmits bankruptcy notices for every federal bankruptcy court in the country — is down, stopped by a cybersecurity incident at the contractor that runs it for the judiciary.
What stopped, and what didn’t
On October 15, the Administrative Office of the U.S. Courts confirmed what courts had been telling filers for days: the information technology and websites that BAE Systems operates in support of the Bankruptcy Noticing Center “have been impacted by a cybersecurity incident,” and “the incident has also halted the BNC’s production and distribution of bankruptcy notices for bankruptcy courts.” The unavailable systems include the center’s public websites — bnc.uscourts.gov, bankruptcynotices.uscourts.gov, and notices.uscourts.gov — and the National Creditor Registration Service. According to the U.S. Bankruptcy Court for the Eastern District of Texas, BAE first told the Administrative Office about the disruption late on October 9; both electronic and paper notices are affected nationwide.
Bankruptcy courts continue to accept filings, open cases, and process documents. The Notices of Electronic Filing that the courts’ own filing system sends to registered attorneys and trustees have continued. What is missing is the centralized noticing layer: the official notices the center generates and delivers to creditors, debtors, and parties in every case. A BAE Systems spokesperson said the company is “aware of a security incident that is isolated to the BNC system,” that it took immediate steps to address it, and that it is coordinating closely with the Administrative Office; BAE says it is investigating and working to restore operations. Neither the judiciary nor the company has described the cause of the incident, said whether any data was accessed, or announced a date for restoration.
Why a noticing pipeline is the archetype
A bankruptcy notice is not an announcement. It is a legal instrument with deadlines attached. The meeting-of-creditors notice tells parties where a case begins; claim deadlines and discharge orders move through service of notice, and a party who is never notified may not be bound by the clock everyone else is running on. When the pipe stopped, the clocks did not stop with it — so judges have had to stop them by hand, district by district. The chief bankruptcy judge for the Eastern District of New York ordered that no party will be penalized or lose legal rights for missing a deadline or failing to respond because of the outage. The Eastern District of Texas extended automatic-stay protection by thirty days for a class of repeat personal bankruptcy cases. Other districts are handling outage problems case by case.
There is no second noticing system waiting in reserve. The fallback is judicial improvisation — orders, extensions, and warnings to check the docket yourself. Even the self-help route runs through the same failure: parties who would serve documents themselves normally generate creditor mailing lists through the National Creditor Registration Service, one of the systems that is down. The Eastern District of Texas also reports that the disruption interrupted the court polling and certification processes that support notice delivery. One producer, one pipe, statutory deadlines on the output, and no substitute channel: that is the single-point-of-failure archetype in its purest form, and it was sitting inside a routine-looking service contract.
The questions a sole path has to answer
Any organization can find its own version of this dependency: a function with one producer, no alternate route, and consequences that compound daily. The plan for such a function has to answer four questions before the outage, not during it:
- What does the contract promise on the day the vendor stops? Continuity and failover obligations, recovery-time commitments, and the duty to notify and cooperate are either written into the agreement or they are assumptions. An assumption discovered mid-outage is not a plan term.
- Could anyone else run the job with what is on hand? The data, address lists, output formats, and production capacity to print, send, and certify must exist outside the vendor’s walls — escrowed, current, and usable — or the honest answer is that no one can run it, at any price, on day one.
- What is the manual fallback, and when was it last rehearsed? A fallback that has never been exercised is a hope with a document number. The fallback also has to cover its own dependencies: if the workaround needs data or lists from the failed system, it is not a workaround.
- Which outputs get triaged first when the pipe stops? Not everything a sole-path function produces is equally load-bearing. Rank the output in advance — which items carry legal deadlines, which can wait a week — so day one is spent executing a priority order instead of inventing one.
This desk made the same point from the other direction in Your Plan Names the Vendor. Has It Ever Survived the Vendor Failing? — a plan that lists a dependency is an inventory, not a capability. The bankruptcy courts are demonstrating the corollary at national scale: when the dependency is the only path, the inventory is all there is, and the people downstream start improvising the capability in public.
The lesson for your own vendor map
The generalizable lesson is that the dangerous line on a vendor map is rarely the biggest vendor or the most expensive one. It is the one whose output somebody else’s deadlines depend on, produced by exactly one pipe, with no rehearsed way to make that output another way. That dependency usually looks like administration — noticing, invoicing, credentialing, reporting — right up until the day it stops and the deadlines keep moving.
So the question to take back to your own plan is not whether your vendors are reliable. It is narrower and harder: for each critical output, if the sole producer went silent tonight, what would you be doing on day ten — executing a fallback you have run before, or writing your version of an emergency order, one deadline at a time?