Which dates, time ranges, units and existing monitoring records are needed to describe a persistent change, without treating one cloudy day or one screen value as proof of failure?
1. Build a dated record, not a single-point conclusion
Use records already available to you: monitoring history, utility statements if relevant, saved alerts, commissioning paperwork, and service records. For each entry, preserve the exact date or time window, the displayed quantity and unit, where the reading came from, and whether the interval is complete. Mark missing periods as missing; do not fill them with estimates. Keep the screen’s label as shown, and ask whether it represents system production, exported energy, or battery flow. Those quantities are not interchangeable just because they appear together in an app or bill.
Separate one unusual day or isolated screen value from a change that recurs across comparable, documented periods. A utility bill or battery display alone cannot establish how much the array produced or why a value changed. If the available records are incomplete, say so; a gap is not evidence of zero production.
How do location, weather, system configuration and age affect what can be compared, and which expected-output baseline belongs to this installation?
2. Compare against this system’s own documented basis
There is no single daily output figure that fits every home. Location, daily weather, system configuration, module age, and known changes can affect what periods are comparable. Record context you already know, such as a system change or weather event, without treating it as the cause. Do not go onto the roof or inspect, clean, or move anything to investigate.
Ask the installer or manufacturer for the expected-output baseline for this specific installation and the assumptions behind it: the system model/configuration, comparison period, and the weather or resource conditions used. Ask how the available monitoring records align with that baseline and whether gaps or estimated readings limit the comparison. A federal facility operations guide discusses tracking production over time and comparing measured production with a system model under resource and temperature context. That is useful context for asking what basis was used; it is not a residential target, fault threshold, or promise for your home.
What exact alert text, system identifiers, utility status and contract or commissioning records can be shared with the installer without opening equipment or changing settings?
3. Send the exact records and identifiers safely available
Share copies or screenshots you already have, without opening equipment or changing settings. Include the exact alert wording and date, the monitoring-system or inverter identifier shown in existing records, and the production history with its displayed units and missing intervals. Add commissioning or service documents and warranty or contract records if available. Note the utility’s status only as it is shown in its own account or notice; do not infer array production from it. If an identifier or record is not available, list it as unknown rather than trying to retrieve it from equipment.
Ask the installer to distinguish measured from estimated values, confirm which quantity and system period are being compared, and state whether the comparison is complete enough to interpret. Request the baseline and assumptions in writing. These details help frame a review; they do not establish a fault or predict an output level.
When does the record support a qualified installer/manufacturer review, and which visible damage, heat, smoke, water or electrical warning requires staying clear and emergency response first?
4. Decide whether the pattern needs qualified review
Ask for installer or manufacturer review when dated records show a recurring change across comparable periods, when an exact alert remains unexplained, or when the baseline, units, or missing intervals prevent a fair comparison. Send the records and ask what they can conclude from them and what information is still needed. One low reading, a bill, a battery-flow display, or weather by itself does not identify a cause.
Source context checked October 15, 2026: U.S. Department of Energy, Federal Energy Management Program, “Life Cycle of Photovoltaic Systems: Operate and Maintain an Existing Photovoltaic System”, sections “Confirm PV System Performance” and operational indicators. Publication or page-update date not established. This is federal-facility O&M guidance, not a residential diagnostic threshold or guarantee.