1083680 - [Usage][Cost Control] (Follow up to fix root cause) Noticeable difference between by application breakdown totals and the total displayed in chart and widget

Status: RESOLVED WONTFIX

(Firefox OS Graveyard :: Gaia::Cost Control, defect, P1)

Description

[Marina Rodríguez [:mai]]

+++ This bug was initially created as a clone of Bug #1080473 +++
The bug 1084473 is a front-end workaround, when this bug was resolved on the backend, the workaround must be removed.

[Blocking Requested - why for this release]:
The user could get confused for the "missed" mobile usage. Furthermore, the difference is huge pointing to a possible problem in the platform.

STR:
 1. Make reset-gaia 
 2. Don't enable any mobile data network.
 3. Launch Usage and pass the FTU with default values.
 4. Obvserve usage is 0
 5. Press home button
 6. Enable 3G from utility tray
 7. Keep home pressed and return to Usage application
 8. Wait for chart and application breakdown

Expected:
 Total amount of data displayed in the breakdown sums up the same amount displayed in the chart.

Actual:
  There is a noticeable difference between the two totals.

Comment 1

[Marina Rodríguez [:mai]]

After "Bug 1077546 - [Cost Control] Add app detail view" was landed, the difference between by application breakdown totals and the total by interface (residual traffic) is visible on the system App detail view. 

Now, the amount of mobile traffic of the System app that is showed on the app list differs from the amount on the app detail view.

This issue should be resolved when the residual traffic was 0.

Comment 2

[Salvador de la Puente González [:salva]]

In the NetworkStats API, both types of data-usage queries (by app and global) should return the same totals but it seems some implementation details prevent us from getting this working as expected. We should address the back-end metering problem or try to reduce the difference as much as possible.

Current solution in bug 1080473 is accumulating the difference (global - by-app) in System APP but this is like moving other applications' traffic to System.

Comment 3

[Wesley Huang [:wesley_huang] (EPM) (NI me)]

triage: blocking

Comment 4

[Salvador de la Puente González [:salva]]

Assigning to :mai just to notice we are looking over it. As said in comment 2, the problem is solved as work around but a more accurate solution is needed.

Comment 5

[Kai-Chih Hu [:khu]]

Do we have any update for this bug? Thank you.

Comment 6

[Marina Rodríguez [:mai]]

Nothing, the origin of the residual traffic still be a mystery

Comment 7

[Francisco Jordano [:arcturus] [:francisco]]

Marina, do we have an estimate for this bug?

Comment 8

[Marina Rodríguez [:mai]]

Hi,
this bug has to be resolved on the backend, currently we are using this workaround to distinguish the browser traffic from system traffic.

Nowadays, the principal consequence of this bug, it is the app doesn't show the system traffic graphic because we don't know how to show the residual traffic. 

Best regards.

Comment 9

[Ken Chang[:kenkai]]

Hi Ethan, I wonder if this is a new feature for network metering. If we want to have this, is it possible to finish this in 2.2.

Comment 10

[Ethan Tseng [:ethan]]

(In reply to Ken Chang[:ken](OOO from 2/18 to 3/1) from comment #9)
> Hi Ethan, I wonder if this is a new feature for network metering. If we want
> to have this, is it possible to finish this in 2.2.

This is not a new feature. It's an obvious but troublesome bug of NetworkStats component (which is the backend Marina meant in comment 8). Platform engineers discussed this issue several times, but so far we have not come up with a viable solution yet.

Bug 1102812 was filed for the same purpose as this bug. We had discussions with Necko team members in that bug, and I'll keep tracking this issue there.

To resolve this issue, more investigation is required before we work out a concrete solution. It's difficult to estimate the schedule right now.
When is the deadline of 2.2?

Comment 11

[Ken Chang[:kenkai]]

[Tracking Requested - why for this release]:

Currently, we don't have a good solution and need to discuss more details. And it isn't a regression. I nominate this bug as a tracking-b2g+ bug but not a 2.2+ bug.

Comment 12

[Wesley Huang [:wesley_huang] (EPM) (NI me)]

[Tracking Requested - why for this release]:

Comment 13

[BMO Automation]

Firefox OS is not being worked on