Missing Days in Outputs and PV Ladder


#1

Something seems to be amiss in the aggregation of past outputs into the PV Ladder results and possibly other metrics.
My system seems to be going backwards in lifetime energy generation, due to the number of days being counted reducing over time.
A few days ago, the system’s lifetime stats as shows on PVLadder showed like this:


This morning, I see the system has dropped 140 spots down the list (usually it rises by ~5 - ~30), and shows like this:

The lifetime generation has gone backwards, the ‘outputs days’ has lost 4 days rather than gain a few, and the health % has dropped, indicating a larger number of days ‘missing’ between today and the start date of the system being logged.
My Fronius inverter meanwhile, which is the source of the generation stats and has been logging since the day it was installed, indicates lifetime generation is 9.41 MWh, whereas in the early days it was in agreement with the PVOutput lifetime energy generation, which you’d expect since both are tracking the same source data.
PVOutput seems to be progressively ‘evaporating’ days when calculating the lifetime totals, yet the daily table seems to go all the way back to mid-May when the system was first established, 331 days ago now, so as of today it seems 35 days of ‘outputs’ are being ignored.

(I first noticed a large ‘backwards jump’ and the health% drop a month or two ago, but only now had the screenshot data to prove to myself I wasn’t dreaming. It might be something that only occurs over a weekend, possibly a background cleanup process?, as it seems to work normally over during the week, but I recall the last time thinking ‘thats odd’ also being a Monday morning check on what happened over the weekend).


#2

It looks correct today -

It may have been a cached result, ladder data is calculated every 24 hours. You can also force refresh of lifetime ladder data from the Edit System page.


#3

Yes - I did hit the ‘Recalculate’ button while looking at this yesterday, but it didn’t seem to do anything at the time. Yet Lo, this morning, the system (sid=57175) has jumped up 870-odd places, the outputs count is back to the full 332, and health% back to 100%, all fixed, after being off for several months.

Wierdly, so have the other three systems, and I didn’t touch the ‘Recalculate’ button for them at all.

It wasn’t a browser-cache thing, as the output was identical on several devices, laptops and tablets.

Maybe it read the post, heard me swearing and cursing yesterday, and decided to fix itself up automagically. Wierd.