Last updated: September 15, 2026. By Dan McKenna, founder, Elite Energy Management.
Usually, yes, because compressors rarely die silently. In the weeks before one quits, the data your thermostats or energy management system already collect starts to show it: the unit cycles on and off more often as protective trips interrupt the compressor, a stage stops responding when it is called, runtimes stretch, and the space cools more slowly per minute of runtime at the same outdoor conditions. Monitoring that reads that data flags the pattern before the failure becomes an emergency call.
What does a failing compressor look like in the data?
The compressor is the most expensive part of a rooftop unit and the one whose failure is hardest to miss on a hot afternoon, but its decline is visible long before that afternoon:
| Symptom | What it looks like in monitoring data |
|---|---|
| Protective trips | Short cycling: the compressor starts, runs briefly, shuts down, and repeats, far more often than the unit's own history |
| A stage that no longer responds | The thermostat calls for cooling and the expected stage never engages, or engages late |
| Weakening output | The space cools more slowly per minute of runtime at the same outdoor conditions, so cycles run longer to remove the same heat |
| Comfort that holds, then breaks | Setpoint is maintained early, then slips on the hottest days, which is usually when someone finally calls |
Every one of those is a comparison against the unit's own weather-normalized baseline, not a fixed threshold. A unit running long during a heat wave is normal; a unit short cycling at 85 degrees when it ran smoothly at 85 degrees a month ago is not. The methodology page covers how those comparisons are made.
Why don't you find out sooner?
Because the equipment compensates. A rooftop unit with a struggling compressor keeps hitting setpoint by running longer, so the thermostat has no reason to alarm, and nobody in the building feels anything until the day the compressor cannot keep up. By then the failure is an emergency service call at 2 to 5 times the planned cost, and a compressor that has been tripping on overload for weeks is often past saving. A planned repair while the unit is still cooling can run in the hundreds; a compressor replacement after a full failure regularly cascades past $5,000 once labor, refrigerant, and downtime are counted.
How much warning do you get?
Enough to schedule instead of scramble. Across technician-confirmed repairs on the monitored fleet, all fault types included, 87% were flagged before the technician arrived, with a median of 6 days advance warning. Six days is the difference between getting a quote, deciding on repair versus replacement, and booking a planned visit at standard rates, versus paying after-hours pricing while a dining room heats up.
What can't the data tell you?
Plenty, and it is worth being clear about. Monitoring shows that a unit is behaving the way compressors behave when they are failing; it does not confirm the compressor is the cause, quantify how much life is left, or tell you whether to repair or replace. A failing contactor, a refrigerant problem, or an electrical fault can produce overlapping signatures, and the technician on site sorts that out. What monitoring changes is when that technician arrives and what they arrive knowing. The companion post on refrigerant loss covers the most common overlapping signature.
Frequently asked questions
Can software tell a failing compressor from a refrigerant leak?
Often, but not with certainty. Short cycling and a stage that stops responding point toward the compressor or its electrical side, while a gradual loss of cooling output with normal cycling points toward refrigerant. The platform reads both patterns and says which it is seeing, and the technician confirms on site. Either way the unit needed attention, and the early flag is the value.
Does this require current or power sensors on the unit?
No. The signatures above come from temperature, staging, and runtime data that connected thermostats and energy management systems already produce. Where power metering happens to exist at a site, it adds a second signal, but it is not required.
What should I do when a unit gets flagged?
Schedule a planned visit while the unit is still cooling, and send the technician the flag and the data behind it so they arrive with a hypothesis. Do not wait for the comfort complaint; the whole point of the warning is to spend it.
What does this cost?
Elite Energy Management runs $5 to $12 per rooftop unit per month on standard tiers, with no hardware and no installation, full pricing here. One compressor caught before it fails pays for years of monitoring on that unit.
If your locations already have an energy management system or connected thermostats, a free 30-day pilot will show you which rooftop units are working harder than they should be: read-only access to your existing systems, first health scores within days, no hardware, no commitment.
Related reading: how to tell whether an RTU is losing refrigerant remotely, how detection and validation work, and a plain-English guide to fault detection and diagnostics.