As IT leaders, most of us plan for inclement weather. We think about generators, UPS capacity, backups, and keeping critical systems online if the power goes out. Those are the obvious risks. The blackout scenarios.

But the events that cause the most operational surprises are often the brownouts, not the blackouts.

A facility can still have power while systems slowly degrade. Voltage fluctuations, partial utility failures, or unstable grid conditions can create situations where equipment technically remains online but does not behave normally. Monitoring may show everything as “up” while cooling struggles, airflow drops, or power conditioning begins stressing hardware.

Cooling is one of the first things I think about during severe weather planning. If you cannot physically reach the building because of flooding, ice, or storm damage, do you have visibility into whether your AC units are operating properly? Is airflow being monitored, or just temperature? Many environments assume that if power exists, cooling exists. That assumption can fail quickly.

Then there are the operational dependencies people rarely think about until something breaks.

Security cameras. VoIP phones used by elevators. Fire alarm panels that depend on network connectivity. Door access systems. Edge switches feeding safety infrastructure.

If your primary uplink fails during a storm event, who is responsible for implementing the network changes needed to maintain connectivity? Do you have documented procedures for modifying static routes or activating backup circuits remotely? Do you know who has the authority to initiate a controlled shutdown if environmental conditions start trending in the wrong direction?

Severe weather planning is not just about infrastructure. It is about roles and decisions under pressure.

One of the most useful things a team can do is simulate these events before they happen. War room it. Pick three realistic scenarios and walk through them as if they were unfolding in real time.

What happens if your primary circuit drops during a storm? What happens if cooling becomes unstable but the power stays on? What happens if the building becomes inaccessible for multiple days?

Run the scenario. Document the response. Then regroup quickly afterward while everything is still fresh and capture what was learned.

One of the biggest surprises I experienced during a flooding event had nothing to do with servers or networks.

We needed to print checks.

Because our normal process depended on systems and materials that were not accessible during the incident, we had to solve something we had never planned for. That experience permanently changed how I think about continuity planning.

Technology outages rarely fail exactly where you expect them to.

So I am curious how others approach this.

What unexpected issue surfaced during a real event that you had never planned for?