Pull a quarter's worth of alarm data off a refrigerated fleet and the first thing you notice is the volume. Across the top 20 fault codes alone, our last 90-day window logged more than 6,000 alarm events. The instinct is to treat that as 6,000 problems — a wall of red that needs a wrench behind every line.

It isn't. And learning to tell the difference between an alarm and a problem is most of the job.

Root cause

Your number-one fault code is a behavior, not a failure

The single most frequent alarm in the dataset, by a wide margin, was Box Temp Out-of-Range — just under 1,500 events. Add its sibling code on the secondary box and you clear 1,600. That's more than a quarter of all the alarm traffic in the top 20.

Here's the thing: almost none of it is a refrigeration failure. The overwhelming cause is doors being open during loading and unloading. The unit is doing exactly what it should — flagging that the box drifted out of its setpoint band — but the root cause is a dock door, a slow loader, or a long dwell, not a sick compressor.

That one fact changes how you read everything underneath it. If you chase box-temp alarms as mechanical events, you'll burn a service budget running down phantom failures and you'll desensitize your team to the alarm that does matter — the one that fires when there's genuinely no cold air. The fix here isn't a part. It's an SOP: tighten loading discipline, set realistic dwell expectations, and route box-temp alarms to operations first, maintenance second.

Triage starts with muting the noise so you can actually hear the signal.

Triage

Once you filter the noise, the signal clusters

Strip out the door-driven temperature alarms and the rest of the list sorts itself into two useful piles.

The cheap, proactive fixes. A surprising share of the remaining volume is electrical nuisance codes — Check Starter Solenoid Circuit and Check Starter Circuit together logged roughly 850 events. These almost always trace back to a loose or corroded terminal connection at the starter. None of them is an emergency. But every one is a coin-flip away from becoming a genuine no-start on a cold morning with a loaded box. The smart move is a scheduled, low-cost intervention: clean the terminals, install a fresh pigtail at the starter, and watch those codes evaporate. You're not reacting to a failure; you're retiring a category of future failures for the price of a connector.

The signals that actually move cold chain. The codes worth real attention are the ones tied to refrigeration itself — Check Refrigeration System, Low Refrigerant Pressure, Check SV1 Circuit. Lumped together that's several hundred events, and the most common culprit is simply low Freon. Unlike an open door, these don't self-resolve. A unit slowly losing charge will keep alarming, keep struggling to hold setpoint, and eventually put product at risk. These belong at the front of the queue.

Concentration

A few units do most of the screaming

The other pattern worth naming is concentration. Alarms are not spread evenly across the fleet — they pool around a handful of problem children. Our worst-offending unit threw over 300 alarms in 90 days, nearly double the next unit on the list, and the top few assets accounted for a hugely disproportionate slice of total volume.

This is the classic maintenance Pareto, and it's good news, because it makes the work finite. You don't have to fix the fleet. You have to fix the five or six units generating the noise.

Pull each one's alarm profile and the action usually writes itself: a unit dominated by refrigeration codes gets flagged for a system review and a charge check; one buried in alternator and charging alarms gets a battery-and-alternator workup or a dealer assessment for a persistent electrical fault; one lit up with fuel-sensor resistance spikes gets the sensor circuit inspected for corrosion before it strands a driver.

Playbook

What 90 days of data is really for

The point of an alarm report isn't the count. It's the triage. Three habits do most of the heavy lifting:

• Separate operational alarms from mechanical ones. Door-driven box-temp codes are an ops conversation, not a maintenance ticket. Route them accordingly so your shop isn't chasing ghosts.
• Convert nuisance codes into scheduled work. Loose-terminal starter alarms are free warning shots. Spend the connector now, not the road call later.
• Work the repeat offenders, not the fleet. A small number of units generate most of the noise. Flag them, profile them, and assign each one a specific next action.

Do that, and 6,000 alarms stops being a wall of red and becomes what it actually is: a prioritized to-do list, most of which you can close from a desk.