There was another topic about this but that’s still unsolved.
OpenVPN daemon is always a little more than other apps combined. And that has to be the total data usage, but GW adds other apps usages on top of it. So I think it’s showing more than actually used.
Here’s screenshot: s14.postimg.org/i0esq980h/GW_data_usage.jpg
What do you think?
Yes, in some cases GlassWire can show more traffic with a VPN due to the way a VPN works. We are working on improving VPN data counting.
Same problem here. This makes GlassWire unusable for many of us. Hopefully a solution will be coming our way soon.
I saw this in a duplicate thread:
And I see you here much more recent saying something in similar regards and I was wondering if you were any closer to solving this problem? I run almost exclusively on a VPN +/- a few MBs here and there and my traffic seems almost duplicate most of the time. The problem is, often times, well not that often, but around 3-4 times a month, my VPN doesn’t reconnect after a disconnect and my system runs on my IP. I’m not worried about that since I have a rather extensive IP filtering firewall (which I strongly suspect is responsible for some of the reconnect issues), but it’s times like that where tens to a 1-2 hundred GB are not accounted for by my VPN, so it’s rather tedious to find exact data usage.
I’m really interested in purchasing your software because it seems among the best, since neither DD-WRT nor OpenWRT router firmware monitor is available for my D-Link DIR 655 router. Thus I would really like to know if you are any closer to solving this VPN doubled data usage issue?
Different VPN services work differently, so we’re trying to find a good solution for all VPN users.
Not to be rude, but wow, thank you for telling me what I already know, considering I’ve gone through about a half dozen VPNs in just the last decade. That is, though, not what I asked. I asked if you are any closer to solving the VPN double data usage issue and if so, I would like to know any prediction you may have on if or when it will be available. As I said, I like your software or more specifically its possibilities, but I’m not terribly eager to purchase something that is this dysfunctional for my needs.
Hi
I’ve got the same “problem” - I use OpenVPN Connect. I hope I’ll be done soon.
Best regards
Rafal
Anyone know any counting software that shows the exact correct data counts with all VPN software and networks? If so please give a link to check it and let us know how it works and we’ll see what can be done.
For exact data I only know of router firmware tools like DD-WRT or OpenWRT, and those give about as exact as possible on everything that connects through that router. The only thing is, the router can’t be Ubicron architecture, it has to be like Atheros or Ralink or Broadcom or MediaTek or Qualcomm or just about any other architecture except for Ubicron, which is my problem currently and I really love this Dlink DIR 655 which uses Ubicron.
As for isolating just the VPN, your software actually does that exceptionally, but then it doubles the data usage on the Total. If you want a tool to just monitor the VPN alone, I would recommend Adapter Watch at nirsoft . net (since I’m new I can’t post links, sorry).
With Adapter Watch you can monitor the exact usage of the VPN IP(s) used.
nirsoft has several exceptional monitoring tools.
It sounds like we need to add the ability to let users monitor only specific adapters. It will add a lot of complexity to GlassWire to make it so users choose adapters, when users could just isolate the VPN inside our software and find the data amount that way.
So I guess for now we recommend isolating the VPN traffic under “Usage” and it should be correct. It’s not perfect but since all VPN systems are different that’s all we can recommend currently until we decide to allow users to specify certain adapters.
I’m not really following what you mean about isolating. Under usage the VPN is isolated, and the data is very exact, but what it does is add all the other things running to it as if they weren’t actually running through the VPN, which then effectively just doubles the total. I know for a fact everything is running through the VPN through various other tests, like what is my IP, do i leak, torrent IP identification, DNS test and so on which show that everything is in fact running through the VPN. The problem arises when the VPN isn’t running and the counting is exact, so then I can’t just half the total, because what was running when the VPN wasn’t on is exact.
I’m confused 
as to why you are considering such a major change to GlassWire just to resolve the double-counting issue. It sounds like it will take far too long to introduce such a major change. A simpler solution should be found to resolve this issue much more quickly.
I’m not saying this because I don’t want the ability to specify which adapters to monitor and which not to monitor. I do want and support this very useful feature which you’ve already said that you hoped to add:
Correct me if I’m wrong because I’m not sure that I understand the problem clearly. 
But I do think there needs to be a faster solution than a UI change.
I presume that the double counting occurs because the traffic you are counting includes the real adapters and the virtual adapters. At the moment you are probably counting all traffic so traffic on the virtual adapter is effectively double counting because that traffic also has to go through a real adapter.
Assuming I’ve described the problem correctly then there would seem to be a two-step approach to this:
Remove the double counting by only including traffic on real adapters.
This means focussing on the double-counting problem while avoiding any UI changes.
Many developers have found it is difficult to identify which adapters are real and which are virtual but I know that this was largely fixed with Windows 8 (see Using Virtual Switch Filtering). So it is probably only difficult to identify the adapter types with Windows 7.
Add the ability for users to specify which adapters to include and/or exclude.
I think that it makes more sense to do this at a later date because you would have to make far too many UI changes to implement it properly: database changes to record traffic by adapter; add graph and usage by adapter; allow multiple graphs so each adapter can be viewed separately; allow history to be deleted by adapter; etc.
P.S. Edited to correct spelling/grammar and add hyperlink to virtual switch article.
Just out of speculation, can’t the total just be based on what’s going to and from the router/modem? The reason I ask is because under the network tab it lists my router. I haven’t purchased the software yet, so I don’t know exactly what that entails and from what I can tell in the overview, it just shows the connection, not the activity, but considering there can’t really be more than a couple TAP drivers (adapters) active, when the VPN is on it blocks the WinOS TAP Driver and routes it through the OpenVPN TAP, which explains the doubling, unless you have specific proxy settings for like filesharing, browsing or whatever and you are running two active TAPs to the net, but is there no way to monitor the data after the TAPs, as in the actual data going in and out of the PC to the router or modem for the total usage?
It’s not easy to interact with routers because each router is different, but we’re working on some ways to monitor the entire network. And you’re correct that if we did so then it would solve the problem I think in most cases.
Yes, that was kind of my assumtion, as uneducated as it may be, that it’s due to the several different architectures that are in use for routers/modems, but my hope is that you can somehow just focus on the actual net protocols themselves going in/out of the PC, that are all very standard, of which only telnet,netis-router,rfb,xsan-filesystem,microsoft-ds,sip,ssh,netbios-dgm,http and ms-wbt-server are really used on the open net and of those only like 2-3 are commonly used on a PC level.
