OK to collect anonymous usage data?

I actually went through the pain to do this… It was quite ugly. :confused:

it’s ok for me that this statistics are collected. i also like to see them on a public site, maybe with some graphs in the future.

BT Sync collects “data transferred directly, through relay, size of folders, and number of files synced.”

i think it would be interesting to know how many data is transfered with Syncthing and at some point we also can say something like “more than a petabyte of anonymous data has been synced between BitTorrent Sync users”.

http://blog.bittorrent.com/2013/05/06/bittorrent-sync-crosses-one-petabyte-milestone/

they send the data while the client checks for new versions. maybe it would be good to do that also.

in the settings dialog should be a link to the view of the data, that will be transferred.

i have compiled the newest version from repository and activated it. can’t wait to see my client there.

it would be very cool to have a realtime statistic…

https://data.syncthing.net/, file pull requests on https://github.com/calmh/st-usage-reporting :slight_smile:

Agreed. But I don’t think we want to send data in real time (i.e. very frequently), so best bet to get a consistent view right now is “yesterday”.

We’ll probably want some way to segment the data as well, like clicking on “linux-arm” and the top summary table reflects only nodes with that platform, etc.

Nice. Seems a part of my nodes data is online, too :wink:

btw. “min” makes not really sense, isn’t it? Maybe 5% like 95% makes more sense?

Yeah… 5% would probably be better.

Just so I understand this… I’ve just upgraded to 0.8.15, and in Settings ticked “anonymous statistics”, and I didn’t get this info at all. I don’t mind, myself, but do I only get told what anonymous data is sent if I get hassled by the Syncthing UIto turn it on, but not if I explicitly turn it on myself?

Eh, yeah, that’s exactly how it is… I can see that being a bit annoying. A link in the settings dialog should be added to show the same info.

The problem with additional logging is that, anytime you start logging data the the governments will want a bite of it, send you court orders etc as they please. So if you do it please do it in a way that is not possible to identify anyone anytime anyplace.

Yeah, so for the record the server side of this is up on github, linked above. It doesn’t log the IP of the posting syncthing instance and there’s really nothing of any value (apart from statistical curiosity) in the posted data, so there’s no reason to start thinking law enforcement here.

How to read 5% or 100%?

min was clear as it describes that one node with the minimal amount of that data.

The percentages are “x% of the reporting nodes have a lower value than this”. I.e. in this screenshot, for “Memory Usage”, 5% of the reporting nodes use 9 MiB or less, while 95% use 472 MiB or less. Seen in the other direction, 95% of the reporting nodes use 9 MiB or more and only 5% use more than 472 MiB.

The 5% value can be seen as “practically minimum”, disregarding nodes that have just been setup and don’t do anything useful (i.e. have 0 files and 0 bytes) and likewise 95% is “mostly max” in that there’s only a few weirdos exceeding it. :wink:

1 Like

@generalmanager see v0.9 development for the ram utilisation issue.

I would like it better if Global Discovery could be disabled by default. In my opinion it would be better to ask the user (on first start) if it would like to use Global Discovery to connect over WAN (in non-technical language, image, etc.). I don’t like this option because Syncthing does know my newly created setup exists, and ‘data’ has already being send through the Syncthing-server(s).

I’m also voting for a wizard that helps users creating a setup, but this isn’t that imported now. :wink:

Thanks for your great work so far! Really looking out for the 0.9.x branch!