Software that the US Department of Agriculture (USDA) uses to manage loans and payments for American farmers has been going down repeatedly since Monday, November 16, 2026. An internal email reported by Nextgov/FCW and Government Executive blames a "security incident" serious enough that the department reset every database password it has.
What happened
According to Nextgov/FCW, which cites three people familiar with the matter and an email it viewed, USDA's Cybersecurity and Privacy Operations Center "uncovered a security incident." The email, sent Monday by Chris Rottler, acting assistant chief information officer for USDA's Farm Production and Conservation Business Center, said the department's Information Services Division had to "reset all database passwords in all environments, and this is requiring us to reset all applications causing outages."
A technical bulletin dated Thursday, also viewed by Nextgov/FCW, warned that applications "may continue to experience intermittent outages throughout the week." Some staff also reported problems with the department's VPN late Thursday.
Who is affected
USDA runs a financial safety net for US farmers, including crop insurance and support when bad weather hurts a harvest. Nextgov/FCW reports that many of the tools used for payments between the agency and farmers were affected. Agricultural outlet DTN reported that software used by local Farm Service Agency and Natural Resources Conservation Service offices has had intermittent outages since Monday, disrupting some loan and payment work, and that staff were told to "proceed with caution and don't do any critical work."
What is still unknown: who carried out the intrusion, whether any data was taken, how many farmers are affected, and whether payments have actually been delayed. USDA has not publicly confirmed a breach. A senior USDA official told DTN only that outages sometimes happen because of updates.
Why resetting passwords takes everything down
When defenders suspect an intruder may have stolen credentials, rotating them is the standard first move: old passwords stop working, so stolen ones become useless. The problem is that applications also log in to databases with those same credentials. If those secrets are hard-coded, scattered across config files, or shared between many apps, every rotation turns into a manual hunt, and each app breaks until someone updates it. That is the pattern the internal email describes: resetting the database passwords forced the department to reset the applications as well.
What builders should take from it
- Keep database credentials in a secrets manager, not in code or copied config files.
- Give each application its own credential, so one rotation does not break everything.
- Prefer short-lived or automatically rotated credentials, and practice a full rotation before an incident forces one.
If rotating a secret would take your service offline for days, that is a security weakness in its own right.