Three Outages, One Week: What Outlook, the FAA and Your Bank Have in Common

In five days, email went dark for businesses around the world, every departure in the United States was grounded, and customers at several of the country’s largest banks couldn’t get into their accounts. None of these enterprise network outages started with a cyberattack. Each one started with something small inside a big system, and each one got worse while people tried to work out where the problem actually was.

That second part is the lesson for every IT leader running a multi-location network.

Flowchart of Smartblocks Studio Process on a Dark Interface

Microsoft Outlook: a months-old update meets a busy day

On November 12, Microsoft 365 was disrupted worldwide for nearly 24 hours, and Exchange Online (Outlook mail) was hit hardest. Microsoft’s preliminary report traced it to a bug in a networking gateway update deployed back in July. The bug sat quietly for four months until extra traffic from a Teams failover hit a server fleet that was already missing about 14% of its capacity. At the low point, about 80% of those gateway servers stopped responding.

Restarting the servers didn’t fix it, because servers that came back fell over again as soon as traffic returned. Recovery took traffic limits, automated restarts and a fix rolled out across the gateway fleet. Microsoft’s final report is due November 20.

The takeaway: a software or firmware version can look fine for months and still be the cause. When a vendor announces a bug, the first question is “which of my devices are running that version?” If that answer takes a spreadsheet hunt, you’re already behind.

The FAA: one router, a nationwide ground stop

On Friday, November 13, the FAA stopped all U.S. departures for 4 hours and 21 minutes. Officials said a router installed during routine copper-to-fiber work at a remote site began sending roughly 1,000 connection attempts every four to five seconds. That flood disrupted radar and communications data at 17 of the FAA’s 22 en route centers. Officials have called the router the likely cause, and the investigation is still open.

The symptoms were everywhere. The cause was one device. Until the source was found, the safe move was to stop everything. Once the device was isolated, a router broadcast feature was disabled across the FAA’s roughly 60 routers of that type.

The takeaway: a change at one site can take down shared services everywhere, and the fix often has to go fleet-wide. You need to know every device of that model, where it sits and how it’s configured.

The banks: one bank’s bug, a multi-bank slowdown

On Monday, November 16, Chase customers had trouble logging in, posting transactions and sending Zelle payments. Customers at Wells Fargo, Bank of America and other banks reported problems too, with more than 4,700 Downdetector reports for Chase by about 2 p.m. ET. Later that afternoon, JPMorganChase said it had experienced “a temporary issue related to a software bug in on-premise hardware systems” and that it was not cyber-related. A source familiar with the matter told CNN it was not a cloud problem or an attack. Chase said transactions were back to normal later that day.

The ripple effect is the part to notice. Chase said other banks were affected only when they were party to a transaction with Chase, and called it “not a systemic issue.” Zelle said its own network was running normally the whole time. One institution’s bug in its own hardware looked, from the outside, like half the banking system was down.

The takeaway: a software bug in your own on-premise hardware can look like everyone’s problem. The team that can quickly say “it’s this system, at this site, running this version” restores service and trust fastest. That’s a firmware and change-tracking job as much as a monitoring one.

The common thread: diagnosis is the outage

Look past the headlines and the same pattern shows up three times:

  • Something routine met a load condition. A July gateway update, a newly installed router, a software bug in a bank’s own hardware. Each was routine until real-world load or timing exposed it.
  • One cause, many symptoms. One gateway bug, one router, one bug in on-premise hardware. Thousands of users and locations felt it.
  • Time to find the source set the length of the outage. The fixes themselves were simple once the problem was found.

Few companies run an airspace or a global email platform. But almost every enterprise is living through the same kind of change: carriers retiring copper, new SD-WAN and firewall gear going in, firmware updates rolling out, contractors working at remote sites. The same four questions decide whether your next outage lasts minutes or hours:

  1. Which device is misbehaving, and where is it?
  2. What changed, when, and who changed it?
  3. Which alerts are the cause, and which are just symptoms?
  4. Who owns the fix, and do they already have the evidence?
Dashboard interface showing firmware automation data details

How SmartTile Automate answers those questions

SmartTile is SmartChoice’s network monitoring and automation platform for multi-location enterprises. It watches routers, switches, firewalls, wireless, cloud and carrier circuits from every major vendor, then automates the work that comes after the alert.

  • Firmware tracking across the fleet. SmartTile tracks firmware across your devices, so when a vendor flags a bad release, you can see your exposure in minutes, not after a manual audit.
  • Configuration backups and “what changed?” in one click. Device configs are backed up on a schedule, fingerprinted and compared. When something breaks after a change, the before-and-after is already on file.
  • See the flood, not just the outage. Interface traffic, errors, CPU and memory are stored as history. Set a threshold right on the graph, and a spike on one link becomes an alert with the graph attached.
  • One outage, one ticket. A site that goes dark opens one ticket for the site, not one per device. When a VPN hub drops, SmartTile opens one ticket on the hub and lists every affected site and device. Engineers chase the source instead of dozens of symptoms.
  • Straight to the right engineer. Alerts route by rule into your IT Service Management (ITSM) system (Kaseya Autotask and BMS, ServiceNow, Jira Service Management, HaloPSA and more) with the device, site and readings attached.
  • Carrier fault? Ticket already open. When the problem is a carrier circuit, SmartTile opens the carrier ticket through the carrier’s API, or Joey, our voice agent, calls the carrier. SmartTile has opened 320 carrier tickets since May 2026.

And when the problem is a provider like Microsoft? You can’t fix their servers. But SmartTile helps you prove quickly that your circuits, firewalls and switches are healthy, so your team stops troubleshooting the wrong network and starts communicating with users.

What this looks like on a bad day

A contractor brings a new router online at a branch. SmartTile is already polling that site. The uplink graph spikes and CPU climbs on the neighboring router, crossing the threshold your engineer set. The alert arrives with the graph attached. Devices at a site that loses connectivity are grouped into one ticket for that site. That ticket lands with your senior network team carrying the device, the site and the readings, with the latest config backup and comparison one click away. Your engineer starts at the source.

SmartTile does not claim it would have prevented any of the outages described here, and nothing in this post is a statement about Microsoft’s, the FAA’s or any bank’s systems or vendors. This post is about the capability every enterprise network needs: getting from symptom to source fast.

See your network the way SmartTile sees it

Start a 15-day proof of concept. It’s read-only by default and makes no changes to your devices. Explore SmartTile Automate or see how SmartTile compares.

Sources:

What happens after the alert?

The network goes down. SmartTile detects it, opens the ticket
in the customer’s ITSM system, gets the carrier ticket opened.

Dashboard with data analytics and performance metrics display