
Outlook went down. Then they grounded every flight in the country.
Within two days last week, Microsoft's email system went down worldwide for about a day, and the FAA grounded every flight in the country. The two had nothing to do with each other, but both came down to something ordinary: a routine network change at Microsoft and a routine equipment upgrade at the FAA.
What happened
Microsoft 365 — Outlook, Exchange, Teams, Copilot — started failing worldwide on November 12. Microsoft's own explanation: a flaw introduced during an earlier network upgrade surfaced after a more recent network change. The process that sets up secure connections got overwhelmed, so servers stopped answering requests. No cyberattack, no AI agent gone rogue, just a routine network change. It took about two days to fully untangle. About 430 million people use Microsoft's services, according to a University of Cambridge researcher quoted by Al Jazeera. Problems showed up in more than 70 countries — a Toronto hospital network declared a "code grey" (the alert for a systems failure), some NHS staff in the UK fell back on paper records, courts in the Philippines and Singapore had outages, and the Anderson Economic Group estimated the lost productivity at around $7 billion.
The next day, separately, the FAA ground-stopped every flight in the country for a little over four hours. The cause, per the FAA and Transportation Secretary Sean Duffy: a newly installed router — part of a routine copper-to-fiber upgrade at a radio facility in El Dorado, Arkansas — started firing roughly a thousand connection attempts every four to five seconds and choked the equipment around it. Not a cyberattack either. Investigators still don't know whether the router was faulty or set up wrong, and the FAA has paused installing more of them. Close to 8,000 flights got canceled, another 10,000-plus delayed, and the ripple reached Europe, where airlines held US-bound departures for hours. It's since turned into an argument in Congress over FAA funding, but that's a topic for another post.
The bottom line
A few weeks ago, I wrote about the AWS outage in May, when one air conditioner in a Virginia data center took down Coinbase, FanDuel, and a chunk of the internet for 28 hours. Last week's two outages teach the same lesson.
It's tempting to read this as proof that the cloud is bad, or the government can't run anything. When your email is down, that's how it feels sometimes. The systems that have brought us huge productivity gains and 24-hour connections to everything and everyone are complex and tightly connected, and one person or one part can bring a big piece of them down.
There's no fix for that. You're not going to build your own physical email server (who would it send email to?) or air traffic control system, obviously. What you can control is how your team keeps working when one of them goes down. For a company our size, that comes down to four steps, in this order:
1 - List the two or three services your team can't work without — usually email, phones and your main business app.
2 - Set up a backup way to reach each other that doesn't run on Microsoft 365, like a group text on mobile phones, and keep phone numbers somewhere other than your mailbox. Give your key customers an emergency number to call.
3 - For your main business application, have a manual process to record transactions that are essential to capture during an outage and can be entered later.
4 - Write these steps into your business continuity plan and walk your team through it once.
If Microsoft or your key systems went dark for a day tomorrow, would your team know what to do in the first hour?
These are conversations that we have with Revolution Group clients. Let me know if we can review your business continuity plan with you and help close the gaps.
— Rick