Kivorent News

Wednesday, November 18, 2026

Tech and AINews

Workday says a disaster recovery test hit live databases, causing its outage

Workday says its outage on November 13 and 14 began when backup-site services, during a disaster recovery check, connected to live production databases. Workers at customer employers could not clock in, see pay or request time off for much of a day.

Warden, covering security, outages and the rules for tech
, 3 min read

~17 hoursEmerson College's Workday tenant was unavailable

What to know

  • Workday's status notice says disaster recovery services unexpectedly connected to production databases, leaving data in an inconsistent state.
  • The outage began around 5:15 p.m. Eastern on Friday, November 13; Workday said service was back by 11 a.m. Eastern Saturday for most customers.
  • Workday told customers the problem was not a security incident and has not yet published a technical root cause report.

Workday, the cloud software that many large employers, universities and city governments use for payroll, timekeeping and benefits, went down on the evening of Friday, November 13 and stayed down for many customers until late Saturday morning. Workday's own status notice blames a disaster recovery exercise: while checking its backup site, services there connected to the live production databases instead.

What happened

Customer status pages show the outage starting around 5:15 p.m. Eastern on Friday. Emerson College's page logged that time, according to StatusStack, which tracked dozens of customer notices through the night. Engadget reported that complaints on the outage tracker Downdetector spiked at about 5:19 p.m. Eastern, and Startup Fortune, citing Yahoo News and the San Francisco Chronicle, said Downdetector logged nearly 3,000 reports early on.

Workers could not clock in or out, check pay, request time off or enroll in benefits. Universities reported their student, finance, payroll and HR systems unavailable.

Workday has about 11,500 customer organizations, Engadget reported. Some coverage described all of them as locked out, but no source we read confirmed how many were actually affected, so we have not used a figure for that.

What Workday says caused it

According to a Workday status notice that Washington University in St. Louis relayed and StatusStack quoted, the trouble started while Workday was validating the weekly maintenance window for its disaster recovery tenants. Services in the disaster recovery site unexpectedly connected to the production databases, leaving data in an inconsistent state and causing errors and failed jobs. Workday then put its WD1 production tenants into maintenance mode to repair the data. Georgia Tech's relay of the notice also named a second data center, WD5.

A disaster recovery (DR) site is a standby copy of a service, kept ready in case the main data center fails. Companies test these regularly. The failure Workday describes is the reverse of what a test is meant to catch: the standby systems reached into the live ones.

Workday told the University of Chicago that it was not a security issue, StatusStack reported, and a company spokesperson said the outage had nothing to do with a cybersecurity incident, according to The CyberSec Guru. Workday also apologized in a statement: "We sincerely apologize for the challenges this is causing impacted customers," Startup Fortune reported.

How long it lasted

Recovery came tenant by tenant on Saturday, November 14. Calvin University marked Workday restored at 9:45 a.m. Eastern and Emerson College at 10:27 a.m., about 17 hours after its outage began, according to StatusStack. A Workday spokesperson said service was restored by 11 a.m. Eastern Saturday, with some customers coming back later, The CyberSec Guru reported. Washington University's final updates were posted Saturday evening Central time.

As of Sunday, Workday had not published a technical root cause analysis. Iowa State University told its staff that Workday had identified the root cause and was testing a fix, according to StatusStack.

It was the second long outage at a major work platform in two days. Microsoft 365 email and related services were disrupted for about a day starting November 12, which Microsoft blamed on a network upgrade (our story).

Why it matters

When a payroll and timekeeping system goes down on a Friday evening, hourly workers cannot record shifts and employers cannot run pay or benefits on schedule, so a single vendor's mistake reaches workers at thousands of employers at once. Workday's explanation also shows that disaster recovery testing carries its own risk to live systems. The detailed account Workday owes customers will be a post-mortem; outages caused by a vendor's own maintenance, like N-able's decision to take N-central offline, are a reminder of how much depends on a few cloud providers.

Sources

Internet, outages and security

Written by an AI correspondent under standards set by editor Jacob Knight. Spotted an error? Tell us and we will fix it.