Three outages between Thursday 12 and Saturday 14 November added up to 2,749 minutes, about 46 hours, by our arithmetic: Microsoft 365 for 23 hours 50 minutes by Microsoft’s own record, the FAA’s nationwide ground stop for 4 hours 14 minutes, and Workday’s HR and payroll service for 17 hours 45 minutes. All three fell inside 47 hours 40 minutes, from 16:20 UTC on Thursday to 16:00 UTC on Saturday.
By our arithmetic, the two vendor windows are about 33 and 25 times a 99.9% monthly allowance. A 99.9% month leaves 43.2 minutes of downtime in a 30-day month, against Microsoft’s 1,430 recorded minutes. Workday’s 1,065 minutes compare with the roughly 42 minutes its policy text leaves in November. Whether any customer’s own commitment was missed depends on that customer’s downtime and contract.
Business Pill · Embedded Dependency
A short explainer on embedded dependency: what sits inside the systems you buy. It teaches the general idea only and does not discuss Workday or the outage.
The key insight: As we read it, the number that matters is the overlap of clocks: by our arithmetic, three outages added up to about 46 hours inside a 47-hour-40-minute window, and the two vendor windows were about 33 and 25 times a 99.9% monthly allowance.
Update, 15 November 2026, 17:30 UTC: DHS says the outages are not connected. Homeland Security Secretary Mullin told Fox News that the FAA failure and the Microsoft outage were unrelated: “They are not connected whatsoever.” On Microsoft, Mullin said: “The Microsoft, we knew that almost immediately”. On the aviation fault: “You have a 50-year-old system that just failed”, and “It’s just the first time it’s failed nationwide”.
Mullin said “we need the money from Congress” and that the aim is “to build a full, brand-new system on top of the old system and then be able to take the old system totally offline”. Fox summarised that Mullin says DHS has found no connection between the two incidents and that the government needs to stop funding piecemeal fixes. The quotes are our transcription of the 2-minute clip Fox posted at 16:44 UTC.
As we read it, the DHS answer separates the two clocks in this piece rather than shortening them: Microsoft’s post-incident report is still pending, and Mullin did not name a cause for the FAA fault beyond the system’s age. The Transportation Department’s account, in our ground-stop piece, still says the fault is not determined.
What it shows: the Fox News clip (posted 16:44 UTC, 15 November, about 2 minutes) in which the DHS Secretary answers a Fox correspondent’s question on whether the FAA failure and the Microsoft outage could be connected. It is the primary for the quotes in the box above.
- Microsoft Outage Hit 84 Named Organisations in 12 Countries: by our tally, 84 named organisations in 12 countries said, or were reported by BBC, Reuters and WDR to have said, that the outage affected them, 26 of them health bodies.
- Latent Bugs Explained: Microsoft’s 23h50m vs an 11h Median: how a latent bug, a routine trigger, queues and capacity exhaustion turn into a long outage, from Microsoft, AWS and Google reports; the precedents’ median was 11 hours 10 minutes, by our arithmetic.
Three Clocks, One Window
Microsoft’s final service-health entry, as republished by Cloud Admin Hub, gives the incident window: “Start time: Thursday, November 12, 2026, at 4:20 PM UTC” and “End time: Friday, November 13, 2026, at 4:10 PM UTC”. That is 23 hours 50 minutes, or 1,430 minutes, by our arithmetic. Microsoft’s public posts about the incident ran longer, which we counted in our root-cause piece.
The FAA said on its statements page: “The FAA issued a nationwide ground stop for all U.S. airports following communication issues impacting multiple Air Route Traffic Control Centers”. Transportation Secretary Sean Duffy said the airspace opened at 8:15 p.m. EST, NBC News reported. From the first reports of the stop at 21:01 UTC to 01:15 UTC is 4 hours 14 minutes, by our arithmetic; Reuters described a “nearly five-hour shutdown”.
Emerson College, a Workday customer, says on its status page that Workday “took the system offline at approximately 5:15pm E.T. to perform urgent maintenance”. A Workday spokesperson said service was restored by 11 a.m. EST on Saturday, the Associated Press reported. That is 17 hours 45 minutes, or 1,065 minutes, by our arithmetic, as we set out in our Workday SLA piece. The FAA stop and the Workday outage overlapped for three hours on Friday night, UTC.

What 99.9% Allows
Microsoft’s Service Level Agreement for Microsoft Online Services, October 2026 edition, defines Exchange Online downtime as “Any period of time when users are unable to send or receive email with Outlook Web Access” and states “There is no Scheduled Downtime for this service”. Its credit table pays 25% below 99.9% monthly uptime, 50% below 99% and 100% below 95%.
A 30-day month has 43,200 minutes, so 99.9% leaves 43.2 minutes, 99% leaves 432 minutes and 95% leaves 2,160 minutes, by our arithmetic. The SLA measures downtime “in user-minutes”, so the result differs by customer: a user cut off for all 1,430 recorded minutes would sit at about 96.7% for November, in the 50% credit band.
The Workday policy filed in a City of Palm Desert agenda packet states: “Workday’s Service Availability commitment for a given calendar month is 99.9%”. It also says that maintenance running past the planned window “is considered an Unplanned Outage”.
What Each Side Said Went Wrong
Two of the three accounts tie the outage to network maintenance or upgrade work, and the third to testing. Microsoft’s preliminary root cause cites “traffic shifts during standard network maintenance and upgrades”. Duffy said contractor L3Harris was conducting routine maintenance to switch lines from copper wire to fiber, Reuters reported, and said: “Our switches, our routers, actually got overloaded with data”, per NBC News. Workday placed some systems in maintenance mode “to fix an issue found during testing”, the AP reported; a Workday status-page notice quoted by Washington University IT said tenanted services in the disaster recovery site “unexpectedly connected to Production databases”, UC Today reported.
Microsoft said it has “no indication that the FAA nationwide ground stop is related to the recent service disruption”, NBC News reported. No source we read links the three outages to one another.
The Structural Read
As we read Microsoft’s SLA, the 23 hours 50 minutes is an incident window, not a customer’s downtime: the SLA counts downtime in user-minutes, so each tenant’s uptime, and its credit band, depends on how many of its users were cut off and for how long.
As we read the three accounts, none of the organisations has published a final root cause yet: Microsoft’s post-incident report is pending, Duffy said the FAA fault is not determined, and Workday’s cause wording comes from a status-page notice quoted by a customer.
Microsoft, final service-health entry for MO1550821, 14 November 2026
“This was a technical issue not related to cyber or AI / agents”
Three Implications
FOR IT AND FINANCE TEAMS Microsoft’s SLA says claims must arrive by the end of the month after the incident; for a November incident that is 31 December, by our reading of its example.
FOR CONTRACT OWNERS Microsoft’s credit table pays 25%, 50% or 100% of the monthly fee by uptime band; Workday’s policy text sets a 99.9% monthly commitment, and customer contracts can differ.
FOR OPERATIONS TEAMS The FAA stop and the Workday outage overlapped for three hours on Friday night, UTC, by our arithmetic.
The Business Engineer Lens
This story maps onto the Business Engineer framework Enterprise AI: From Software to Substrate.
The framework says: “The value proposition shifts from features to integration depth.”
As we read it, two days of outages priced that depth from the other side: email, payroll and air traffic all stopped when one layer underneath them failed. The framework predates these outages and does not discuss them.

What Is Not Established
Microsoft’s full post-incident report is not out yet, and the FAA’s root cause is still under investigation; Duffy said: “We haven’t determined what the fault was”. The FAA start time is the first report of the stop, not an FAA timestamp, and Workday’s start is Emerson College’s account. Customer-level downtime, and whether any service credits are claimed or paid, is not known. Customer contracts can differ from the published terms.
Microsoft’s SLA says its outage communications are not a confirmation of missed service levels. We did not contact the companies or the FAA for this piece.
The Bottom Line
Microsoft 365, the FAA’s ground stop and Workday added up to about 46 hours of outage inside two days, by our arithmetic. By our arithmetic, Microsoft’s recorded window is about 33 times a 99.9% monthly allowance and Workday’s about 25 times; customer-level downtime is not known. We covered the FAA stop in FAA Ground Stop: Officials to Check Link to $875M AI Tool.
95,000+ executives read Business Engineer for the AI strategy frameworks cited by ChatGPT, Claude, and Perplexity.
A note on sourcing. Microsoft’s times and root-cause wording come from its final service-health entry as republished by Cloud Admin Hub and PennO365; the credit tiers come from Microsoft’s October 2026 Online Services SLA. The FAA wording is from its statements page; the reopening time is Transportation Secretary Duffy’s, as reported by NBC News, and the 21:01 UTC start is the first report of the stop in our records.
Workday’s times come from Emerson College and a Workday spokesperson via the AP; its SLA text is a copy filed by the City of Palm Desert. All durations, totals and multiples are our arithmetic. We did not contact the companies or the FAA.
Sources: Microsoft service-health entry MO1550821 via Cloud Admin Hub · Microsoft Online Services SLA, October 2026 · FAA statement, 13 Nov 2026 · NBC News live blog, 13 Nov 2026 · Associated Press via ABC10, 14 Nov 2026 · Emerson College IT status page · Workday SLA copy, City of Palm Desert agenda packet









