ResilienceOpinion

Not an Attack Doesn't Mean Not a Security Problem: Lessons From a Week of Outages

November 21, 2026 · Brian Deitch

The last week and a half has felt like everything breaking at once. A Microsoft outage hit Outlook and Teams. An FAA telecom failure caused a nationwide ground stop of about four and a half hours. Banking and airline systems went down. Even the NSA reported a temporary degradation of an internal network. Naturally, the question everyone asked was: is this an attack?

The official answer so far is no. At a House Homeland Security hearing on November 18, Director of National Intelligence Jay Clayton said the outages "are not a coordinated or state-sponsored attack" and described them as "separate service outages associated with maintenance and technical changes." For the FAA, Transportation Secretary Sean Duffy pointed to a newly installed router that overwhelmed communications, while also saying the root cause isn't fully clear yet. Lawmakers from both parties want more briefings, which is reasonable.

I'm not going to second-guess the investigations. I don't have any information beyond what's public, and neither does anyone posting confident theories on social media. But I want to push back on how the question gets framed.

Availability Is One-Third of the Triad

Confidentiality, integrity, availability. We all learned it, and then most of us spent our careers treating the third one as somebody else's job. If a router install can ground flights for hours, that's a security problem whether or not anyone malicious touched it. The impact on people is the same.

And here's the uncomfortable bit: the things that cause big self-inflicted outages are the same things attackers exploit. Changes nobody tested at scale. Single points of failure. Legacy systems nobody fully understands anymore. Rollback plans that exist on paper only.

Why "Is It an Attack?" Is the Wrong First Question

During an outage, the honest answer for the first few hours is usually "we don't know." Teams that jump to "attack" freeze changes and burn time on threat hunting while the real fix is a rollback. Teams that dismiss it as "just a glitch" may miss the one time it really is malicious. Both are failure modes.

The better approach: run outages and security incidents through the same playbook. Restore service, preserve evidence, figure out cause. You don't need to know which one it is to start.

Standard disclosure: I work at Zscaler, a cloud security company, and we've had our own incidents. No vendor gets to claim it never goes down. What you can control is how much one change, one box or one region can take with it.

What To Do

  • Treat change management as a security control. Staged rollouts, canaries and a tested rollback for every significant change. Most of these outages started with a change.
  • Find your single points of failure. The one router, the one identity provider, the one DNS provider, the one SaaS app the business can't live without. Write down what happens when each goes away.
  • Practice failing. Tabletop an outage of your most critical SaaS or network service, with security and ops in the same room.
  • Keep evidence even when you think it's "just a glitch." Logs and configs from the hour before an outage are cheap to keep and impossible to recreate.
  • Have an out-of-band way to talk. When Outlook and Teams are down, how does your incident team coordinate?
  • Be careful with the rumor mill. Outage periods are prime time for phishing ("re-authenticate your Microsoft account") and for disinformation. Communicate early, internally and externally.

Bottom Line

Maybe none of these outages were attacks. Good. But resilience is security, and a week like this is the best free tabletop exercise you'll get all year. Use it.

Sources: Newsmax via Yahoo News on DNI Clayton's testimony, Fox News on lawmakers' concerns, The Hill

outagesresiliencechange-managementfaaincident-responsebusiness-continuityavailabilityzero-trust
All postsDiscuss on LinkedIn