A solar array is the rare home upgrade that can quietly break and keep looking fine from the driveway. Panels don’t rattle when a string goes down, and an inverter that’s derating from heat doesn’t make a sound. The only thing standing between you and weeks of lost production is your monitoring platform — and specifically, the alerts you’ve bothered to understand and act on.
I run a 40-panel roof through two Tesla inverters, and the honest lesson from living with monitoring day to day is this: the app will not babysit your system as thoroughly as you’d assume. Some problems trigger a push notification. Others show up only as a subtle sag in the production graph that no alert will ever flag. Knowing which is which is the whole game.
What Alerts Your Platform Actually Sends
Before you build a routine around alerts, it’s worth being clear-eyed about what each major platform really offers homeowners. Marketing pages blur the line between what the installer sees and what you see, and they’re often very different.
| Platform | Alerts homeowners actually get | What you won’t be alerted to |
|---|---|---|
| Tesla app | System offline / not producing, grid outage events, Powerwall status, firmware updates | Gradual production decline, panel-level issues (Tesla doesn’t expose panel-level data) |
| Enphase App (Enlighten) | System status issues, microinverters not reporting, connectivity loss | Slow soiling losses; some detailed event codes are installer-facing |
| SolarEdge (mySolarEdge) | Inverter status, connectivity; basic system health | Most of the detailed alert engine — string faults, error codes — lives in the installer portal |
| Sense | Custom usage alerts, always-on load changes, device-detection events | Anything on the production side beyond a basic solar CT reading |
| Emporia Vue | Per-circuit usage thresholds you set yourself | Production faults — it’s a consumption monitor, not a solar diagnostic tool |
Two takeaways from that table. First, every production platform is good at telling you when something is fully broken — an inverter offline, a system not reporting. Second, none of them are good at telling you when something is partially broken. That gap is where energy waste lives, and closing it is on you.
The Alerts Worth Acting on Immediately
Not every notification deserves the same response. Here’s how I’d triage the common ones:
| Alert | What it usually means | What to do |
|---|---|---|
| System not producing / inverter offline | Tripped breaker, grid fault, or genuine inverter failure | Check your solar breaker and the inverter’s status lights the same day; call your installer if it doesn’t self-recover |
| Gateway/monitoring not reporting | Wi-Fi or cellular dropout — production may be fine | Reboot your router first; don’t panic, but don’t ignore it for weeks either, because you’re flying blind |
| Microinverter(s) not reporting (Enphase) | One or more panels producing nothing | Note which ones and watch for a day; persistent = installer service call |
| Inverter fault / error code | Grid voltage issues, ground faults, internal component trouble | Photograph the code, check the manufacturer’s code list, contact your installer — don’t just clear it |
| Overheating or derating indication | Inverter throttling output to protect itself | Check ventilation and shade around the unit the same week (more on this below) |
| Grid outage event (Powerwall/battery systems) | You’re on backup power | Shed heavy loads; confirm the battery is carrying the house as expected |
The pattern: connectivity alerts are usually benign, production alerts are usually not, and fault codes always deserve a phone call even when the system appears to recover on its own. Intermittent faults are how expensive failures introduce themselves.
The Alert Nobody Sends: Gradual Production Decline
Here’s the uncomfortable truth about consumer solar apps — none of the big three will notify you that your system produced, say, 12% less this spring than last spring. Soiling, creeping shade from a growing tree, a failing panel dragging down a string: these losses arrive slowly, and slow changes don’t trip threshold-based alerts.
So you have to build this alert yourself, and it takes about two minutes a week:
- Pick a recurring check-in. Once a week, open your app and look at the last seven days of production, not today’s number. Daily figures swing too much with weather to mean anything.
- Compare against the same month last year. Every platform keeps historical data. A clear-sky week in May should land in the same neighborhood as a clear-sky week last May. As a labeled example: if your system historically does around 400 kWh in a good May week and you’re suddenly seeing 340 on similar weather, that’s a real signal, not noise.
- Look at the shape of the daily curve, not just the total. A healthy day is a smooth arc. A notch that appears at the same time every day is shade. A curve that’s flattened on top can be an inverter clipping or derating.
On my own roof, this weekly habit has caught things no alert ever would have — the kind of gradual dip that turns out to be pollen season coating the panels, or a neighbor’s tree encroaching a little more each year. None of it was dramatic. All of it was money.
Heat, Ventilation, and Why Alerts Extend Inverter Life
Your inverter is the component in the system most likely to fail first, and heat is its main enemy. String inverters like the Tesla units on my wall are power electronics working hard in a garage or on an exterior wall, and they respond to high temperatures by derating — deliberately cutting output to protect themselves. That’s energy waste in the short term and thermal stress in the long term.
This is why overheating and derating indications are the most underrated alerts in the whole list. When one fires, the fix is often free:
- Make sure nothing is stored against or above the inverter — garages accumulate clutter, and inverters need clearance to shed heat.
- Check that the unit isn’t in afternoon sun. An inverter mounted on a west-facing exterior wall bakes; shading it helps.
- Clear dust or debris from vents and heatsink fins.
- If it keeps derating in mild weather, that’s not an environment problem — get it inspected.
I won’t pretend to quote you a lifespan number for how much cool operation buys you, because honest answer: it varies by unit and climate. But the qualitative physics is not in dispute — electronics that run cooler last longer, and an inverter that spends every summer afternoon at its thermal limit is aging faster than one that doesn’t. Responding to thermal alerts is the cheapest inverter insurance you can buy.
Consumption Alerts: The Other Half of Energy Waste
Production monitoring tells you what you made; it says nothing about what you wasted. That’s where whole-home energy monitors like Sense or the Emporia Vue earn their spot. Both clamp onto your electrical panel and watch consumption, and both let you set alerts your solar app can’t:
- Usage threshold alerts — get pinged when the house is drawing more than expected, which is how you find the space heater someone left running.
- Always-on load tracking — the baseline your home draws at 3 a.m. Watching that number drift upward over months is how phantom loads get caught.
- Per-circuit alerts (Emporia) — with circuit-level CTs, you can watch specific offenders like an old freezer or a well pump.
For a solar home, the payoff is pairing the two sides: shifting discovered loads into your production window (or into a battery’s off-peak strategy) instead of importing grid power for them. The alert finds the waste; the schedule change captures the value.
A Routine That Actually Works
Alerts are only useful if they land in a routine. Mine is simple: act on push notifications the day they arrive, spend two minutes each week on the production graph and year-over-year comparison, and once a season, physically look at the inverters — vents clear, no new clutter, no error lights. That’s maybe fifteen minutes a month.
The systems that quietly lose a season of production aren’t the ones with bad monitoring. They’re the ones where the monitoring was never read.