Chrome OS network metrics update

Chrome OS network metrics update

Add histogram "Network.Shill.DeviceRemovedEvent" for tracking device
removed events per device (technology) type.

Add histogram "Network.Shill.Cellular.DevicePresenceStatus",
"Network.Shill.Ethernet.DevicePresenceStatus",
"Network.Shill.Wifi.DevicePresenceStatus", and
"Network.Shill.Wimax.DevicePresenceStatus" for tracking the presence
of each network device in the system. The presence status is emitted once
every 3 minutes for each device type.

While here, update the value definitions for NetworkServiceError enum to
match the definitions used in Shill. Also correct the histogram name for
WiMax related histograms.

BUG=chromium:411512, chromium:411510, chromium:402202
Committed: https://crrev.com/7fc5f5fef792ce336e77eabefb0bad9fc4f1be46
Cr-Commit-Position: refs/heads/master@{#296223}

[zqiu1, 2014-09-22 23:24:06]

zqiu@chromium.org changed reviewers:
+ isherman@chromium.org

[Ilya Sherman, 2014-09-22 23:33:16]

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
File tools/metrics/histograms/histograms.xml (left):

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
tools/metrics/histograms/histograms.xml:46976: </enum>
Hmm, why did these change?  Were they previously inaccurate, or did the enum get
re-ordered?  In general it's a Bad Idea (TM) to reorder enumerations associated
with histograms, as the labels are not versioned.  That is, if you change these
labels, you will affect the labels for this histogram across all versions of
Chrome, even though previous versions (presumably) used the previous set of
labels.

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
File tools/metrics/histograms/histograms.xml (right):

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
tools/metrics/histograms/histograms.xml:39595: +</enum>
nit: Please name this "BooleanPresent"

[zqiu1, 2014-09-22 23:52:27]

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
File tools/metrics/histograms/histograms.xml (left):

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
tools/metrics/histograms/histograms.xml:46976: </enum>
On 2014/09/22 23:33:16, Ilya Sherman wrote:
> Hmm, why did these change?  Were they previously inaccurate, or did the enum
get
> re-ordered?  In general it's a Bad Idea (TM) to reorder enumerations
associated
> with histograms, as the labels are not versioned.  That is, if you change
these
> labels, you will affect the labels for this histogram across all versions of
> Chrome, even though previous versions (presumably) used the previous set of
> labels.

This enum have been inaccurate for long time, it was expanded and reordered
without much consideration long time ago.

Fix it now so we can correctly interpret the results.

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
File tools/metrics/histograms/histograms.xml (right):

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
tools/metrics/histograms/histograms.xml:39595: +</enum>
On 2014/09/22 23:33:16, Ilya Sherman wrote:
> nit: Please name this "BooleanPresent"

Will update.

[zqiu1, 2014-09-22 23:56:43]

PTAL

[Ilya Sherman, 2014-09-23 00:02:52]

LGTM, thanks.

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
File tools/metrics/histograms/histograms.xml (left):

https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
tools/metrics/histograms/histograms.xml:46976: </enum>
On 2014/09/22 23:52:27, zqiu1 wrote:
> On 2014/09/22 23:33:16, Ilya Sherman wrote:
> > Hmm, why did these change?  Were they previously inaccurate, or did the enum
> get
> > re-ordered?  In general it's a Bad Idea (TM) to reorder enumerations
> associated
> > with histograms, as the labels are not versioned.  That is, if you change
> these
> > labels, you will affect the labels for this histogram across all versions of
> > Chrome, even though previous versions (presumably) used the previous set of
> > labels.
> 
> This enum have been inaccurate for long time, it was expanded and reordered
> without much consideration long time ago.
> 
> Fix it now so we can correctly interpret the results.

Sad times.  Is there at least a comment in place now that will discourage people
from reorganizing this enum again in the future?

FWIW, you might want to rename the histogram that emits this enum, and mark the
old name as <obsolete>, so that people are less likely to accidentally apply
inaccurate labels to old data.

[zqiu1, 2014-09-23 16:18:20]

On 2014/09/23 00:02:52, Ilya Sherman wrote:
> LGTM, thanks.
> 
>
https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
> File tools/metrics/histograms/histograms.xml (left):
> 
>
https://codereview.chromium.org/589233003/diff/1/tools/metrics/histograms/his...
> tools/metrics/histograms/histograms.xml:46976: </enum>
> On 2014/09/22 23:52:27, zqiu1 wrote:
> > On 2014/09/22 23:33:16, Ilya Sherman wrote:
> > > Hmm, why did these change?  Were they previously inaccurate, or did the
enum
> > get
> > > re-ordered?  In general it's a Bad Idea (TM) to reorder enumerations
> > associated
> > > with histograms, as the labels are not versioned.  That is, if you change
> > these
> > > labels, you will affect the labels for this histogram across all versions
of
> > > Chrome, even though previous versions (presumably) used the previous set
of
> > > labels.
> > 
> > This enum have been inaccurate for long time, it was expanded and reordered
> > without much consideration long time ago.
> > 
> > Fix it now so we can correctly interpret the results.
> 
> Sad times.  Is there at least a comment in place now that will discourage
people
> from reorganizing this enum again in the future?
> 
> FWIW, you might want to rename the histogram that emits this enum, and mark
the
> old name as <obsolete>, so that people are less likely to accidentally apply
> inaccurate labels to old data.

Yes, comments were added to prevent people from reorganizing this enum in the
future.
I am reluctant to rename the histogram, the changes had been done long time ago
in Shill, the old data is probably way too old for us to care about. we are just
updating
the labels to match the data, so we can interpret the existing data.

[zqiu1, 2014-09-23 16:18:29]

The CQ bit was checked by zqiu@chromium.org

[commit-bot: I haz the power, 2014-09-23 16:19:44]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/589233003/20001

[commit-bot: I haz the power, 2014-09-23 20:44:49]

Committed patchset #2 (id:20001) as 12aa1ac5c6828f12a38cad241fe3c3435561c89a

[commit-bot: I haz the power, 2014-09-23 20:45:25]

Patchset 2 (id:??) landed as
https://crrev.com/7fc5f5fef792ce336e77eabefb0bad9fc4f1be46
Cr-Commit-Position: refs/heads/master@{#296223}