Feature request: Normalize graph intervals

Ideas
#1

Today I had the program that uploads status data to PVOutput down for about two hours. My status interval on my system is 5 minutes, so it can be assumed that if data hasn’t come into PVOutput for over 20 minutes, the uploader is offline.

My problem is that when looking at the live graph, I can see my energy generation and energy consumption jump once my program started uploading again. The graph is just strung together and isn’t spaced correctly based on the timestamps. I wish this could be updated to have a more normalized X axis so if the system if offline for a few hours, the points are spaced out based on the time.

One solution would be to just add in the missing intervals, but I really shouldn’t have to do that because the graph should be spaced correctly based on time.

#2

Usually this type of behavior is related to your wifi, cable and your internet provider services. However, there are timing issues when uploading via an API which must be addressed. you will have to provide more information to get a definitive answer. Sometimes it just takes time to get started. It’s been three days since you posted so have things settled down by now?

#3

I had my uploader program offline for two hours on purpose because I was messing with it, so that isn’t the problem.

I was suggesting that when looking at graphs, even if the upload timer are not consistent (maybe you upload every 5 minutes, then 10 minutes, 5, 10, 20, 5, 40), the graph should still show that time going by.

As it is right now, on the live graph or daily, if your uploader program misses a few instants to upload (like mine did for two hours), it looks like everything is fine but, it’s hard to tell how much time has passed if the interval is inconsistent.

I understand that this could be fixed by making sure my uploader program doesn’t miss a single interval, but I think the x axis should be used as the time and have it spaced correctly even if you miss a day or a couple of hours. I mean, if you have data points at 9:00, 9:05, 9:10, then skip to 10:05, it just doesn’t look quite right. It looks like the same amount of time has passed between 9:05 to 9:10 and 9:10 to 10:05.

And for the daily graphs, I had the program offline a week ago for about a week before that because the program to get data from our system was also offline. So there’s about a week where I can’t upload data to PVOutput, because that data doesn’t exist, so also jumping from April 18 to April 27 is frustrating because it looks like a single day has passed between those two data points, when really it was a lot longer than that.

I am really happy with PVOutput, though. Everything has been working smoothly with uploading and viewing energy generation data for the past year is really nice.

#4

I see what you are saying now. The gaps in the data are compressed on the X axis and the times not shown. The table below the graph does show the missing points as "-"s but that is a little inconvenient. when you want to see the actual data on a continuous line including the gaps in the data. If you have one datum that is continuously reported you will see the true timeline and the gaps in the data.
There is another peeve of mine in the plotting when there are missing data during every other upload. In this case the plot shows no data for all the times except when the mouse pointer hovers over the missing data time then you see a haloed dot for the data that is present. I put in a suggestion a couple of years ago requesting that these isolated points be plotted as dots but the idea was completely ignored. I don’t know what plotting method he uses but PVOutput may have no way of controlling that part of the graphic display. The only time you see isolated points is when you disable a curve or enable it. during the change you can see the dots but only momentarily. So it is possible to display them but they are quickly removed.