Graph shape appears a little odd

[OneCD]

Hello,

The graph for my primary server shows valleys that seem to indicate zero-byte days:

My backup servers are rarely used (the graphs for each one are almost a flat line and there are no corresponding peaks that match the valleys shown above).

I think the shape should look more like this GIMP'd mockup:


Or, have I misunderstood this graph?

Also, the graph indicates days where I exceeded the mid-range of 93.1GB. I have a slow 8Mb/s link and on a really perfect day, I can download ~ 70GB in 24 hours. Is the difference due to the awesome sabyenc module?

Thank you.

[sander]

AFAIK, the graph means that your downloads on day 9 and 12 and 16 were ... 0 MB. So you cannot truthfully create your gimp graph from that (which shows a download of 90GB on day 9).
IMHO a better graph would be a bar chart, so horizontal line the days, with vertical a bar for the download on that day.

"93.1GB. I have a slow 8Mb/s " ... 8 Mb/s = 1 MB/s. With 86400 seconds in a day, you could download 86400 MB = 86 GB per day. So less than 10% extra ... can your line be 9 Mb/s?

[OneCD]

AFAIK, the graph means that your downloads on day 9 and 12 and 16 were ... 0 MB. So you cannot truthfully create your gimp graph from that (which shows a download of 90GB on day 9).

Except that's not the case. I download continuously - no days off. Certainly no drops down to zero.

"93.1GB. I have a slow 8Mb/s " ... 8 Mb/s = 1 MB/s. With 86400 seconds in a day, you could download 86400 MB = 86 GB per day. So less than 10% extra ... can your line be 9 Mb/s?

8Mb/s is the maximum line-speed. But then, I assume some of that gets consumed with protocol overheads. I can download (at best) around 70GB per day.

Here's my usage shown in DD-WRT:

[sander]

So ... overshoot and undershoot?

You can check the values: in Config -> Server, do View Source, and find "serverData[0]" ... which data do you see for this month?

[OneCD]

So ... overshoot and undershoot?

Yes, that's the best way of describing it.

Here's the figures for September:

"2018-09-01": 52461309985
"2018-09-02": 122218246298
"2018-09-03": 50809504586
"2018-09-04": 60778472807
"2018-09-05": 73869088851
"2018-09-06": 52928925796
"2018-09-07": 14948118100
"2018-09-08": 117224349607
"2018-09-09": 932092
"2018-09-10": 99037106005
"2018-09-11": 104555351540
"2018-09-12": 782336
"2018-09-13": 141164127318
"2018-09-14": 61036801042
"2018-09-15": 68334119746
"2018-09-16": 512000
"2018-09-17": 72185495515
"2018-09-18": 131712995762
"2018-09-19": 66396274996
"2018-09-20": 992987
"2018-09-21": 114928222112
"2018-09-22": 495616
"2018-09-23": 67059254694

[sander]

(Note: if you have more than one server, apart from"serverData[0]" there will "serverData[1]" and "serverData[2]" etc)

Your dates with your highest GB's:
"2018-09-13": 131
"2018-09-18": 122
"2018-09-02": 113
"2018-09-08": 109
"2018-09-21": 107

So ... impossible on your line?

I don't know why. Even if SAB includes non-downloaded par2-files, it is too high.

[OneCD]

So ... impossible on your line?

Yes.

I don't know why. Even if SAB includes non-downloaded par2-files, it is too high.

Agree.

Maybe the day counts are not resetting at a consistent time of day?

[sander]

Maybe the day counts are not resetting at a consistent time of day?

Interesting hypothesis. Based on your SAB values, your mean value per day over those 23 days is 59 GB. That does fit on your 8 Mbps line, doesn't it?

Your DD-WRT graph shows an amount of 65 - 72 GB each day. Those numbers fit on your 8Mbps-line.

So ... hypothesis is that SAB 'chooses' the measure point not at the same moment? Maybe only after a download. Or the measurement is each day at the same moment, but SAB only updates the counter after a download?
Do you do downloads of 20-30+GB each?

[safihre]

The code measures every single character that is received independent from the content or the NZB's, so also the overhead.

But looking at it, the reset code looks really strange.. Maybe I just don't understand it yet, have to double check.
How often do you restart the system the sabnzbd is run on or how often is sabnzbd restarted?
https://github.com/sabnzbd/sabnzbd/blob ... psmeter.py

[OneCD]

Interesting hypothesis. Based on your SAB values, your mean value per day over those 23 days is 59 GB. That does fit on your 8 Mbps line, doesn't it?

Your DD-WRT graph shows an amount of 65 - 72 GB each day. Those numbers fit.

Agree.

So ... hypothesis is that SAB 'chooses' the measure point not at the same moment? Maybe only after a download. Or the measurement is each day at the same moment, but SAB only updates the counter after a download?
Do you do downloads of 20-30+GB each?

Sure did, I’ve been grabbing a few remuxes lately. They’re anywhere from 20GB to 50GB. A 50GB takes most of the day.

I was thinking the same thing. Maybe the counts only update when a download completes.

How often do you restart the system the sabnzbd is run on or how often is sabnzbd restarted?

System reboot is every 2 months (long uptimes make me nervous). I don’t think I’ve restarted SAB since the day 2.3.5 was released.