Add netconfigs for different network type

Add netconfigs for different network type

BUG=catapult:#2584

Committed: https://chromium.googlesource.com/external/github.com/catapult-project/catapult/+/bbe2a3f0c1783720b38dbc7a3aaf2b46706a166e

[nednguyen, 2016-09-29 15:03:58]

nednguyen@google.com changed reviewers:
+ kouhei@chromium.org, pmeenan@chromium.org

[nednguyen, 2016-09-29 15:03:59]

https://codereview.chromium.org/2377373002/diff/1/telemetry/telemetry/page/tr...
File telemetry/telemetry/page/traffic_setting.py (right):

https://codereview.chromium.org/2377373002/diff/1/telemetry/telemetry/page/tr...
telemetry/telemetry/page/traffic_setting.py:23: NETWORK_CONFIGS = {
Patrick: you know which is the best sources to get these numbers?

[nednguyen, 2016-09-29 15:04:04]

The CQ bit was checked by nednguyen@google.com to run a CQ dry run

[commit-bot: I haz the power, 2016-09-29 15:04:12]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-29 15:26:42]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-29 15:26:43]

Dry run: This issue passed the CQ dry run.

[Pat Meenan, 2016-09-29 19:45:45]

https://codereview.chromium.org/2377373002/diff/1/telemetry/telemetry/page/tr...
File telemetry/telemetry/page/traffic_setting.py (right):

https://codereview.chromium.org/2377373002/diff/1/telemetry/telemetry/page/tr...
telemetry/telemetry/page/traffic_setting.py:23: NETWORK_CONFIGS = {
On 2016/09/29 15:03:59, nednguyen wrote:
> Patrick: you know which is the best sources to get these numbers?

Probably best to use the same presets as dev tools:
https://cs.chromium.org/chromium/src/third_party/WebKit/Source/devtools/front...

[nednguyen, 2016-09-30 01:28:00]

https://codereview.chromium.org/2377373002/diff/1/telemetry/telemetry/page/tr...
File telemetry/telemetry/page/traffic_setting.py (right):

https://codereview.chromium.org/2377373002/diff/1/telemetry/telemetry/page/tr...
telemetry/telemetry/page/traffic_setting.py:23: NETWORK_CONFIGS = {
On 2016/09/29 19:45:45, Pat Meenan wrote:
> On 2016/09/29 15:03:59, nednguyen wrote:
> > Patrick: you know which is the best sources to get these numbers?
> 
> Probably best to use the same presets as dev tools:
>
https://cs.chromium.org/chromium/src/third_party/WebKit/Source/devtools/front...

Done.

[nednguyen, 2016-09-30 01:28:07]

The CQ bit was checked by nednguyen@google.com to run a CQ dry run

[commit-bot: I haz the power, 2016-09-30 01:28:11]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-30 01:54:50]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-30 01:54:51]

Dry run: Try jobs failed on following builders:
  Catapult Windows Tryserver on master.tryserver.client.catapult (JOB_FAILED,
https://build.chromium.org/p/tryserver.client.catapult/builders/Catapult%20Wi...)

[kouhei (in TOK), 2016-09-30 04:20:57]

lgtm

[nednguyen, 2016-09-30 09:50:45]

On 2016/09/30 04:20:57, kouhei wrote:
> lgtm

Patrick: I suspect the ts_proxy's set command is not honored on windows:

  Traceback (most recent call last):
    File
"C:\b\build\slave\catapult\build\catapult\telemetry\telemetry\page\page_run_end_to_end_unittest.py",
line 362, in testTrafficSettings
      latencies_by_page['fast'] + 5)
  AssertionError: 1.246000051498413 not greater than 5.3480000495910645

On my Mac & Linux station, using REGULAR_2G setting make it takes around 10s to
load the page. I would expect similar latency on windows.

[Pat Meenan, 2016-09-30 16:39:52]

On 2016/09/30 09:50:45, nednguyen wrote:
> On 2016/09/30 04:20:57, kouhei wrote:
> > lgtm
> 
> Patrick: I suspect the ts_proxy's set command is not honored on windows:
> 
>   Traceback (most recent call last):
>     File
>
"C:\b\build\slave\catapult\build\catapult\telemetry\telemetry\page\page_run_end_to_end_unittest.py",
> line 362, in testTrafficSettings
>       latencies_by_page['fast'] + 5)
>   AssertionError: 1.246000051498413 not greater than 5.3480000495910645
> 
> On my Mac & Linux station, using REGULAR_2G setting make it takes around 10s
to
> load the page. I would expect similar latency on windows.

I can reproduce the timings locally but I'm trying to understand exactly what is
being tested.  It looks like it is loading file://green_rect.html.  Is there
some logic somewhere in telemetry that converts that into a http page to
localhost?  Otherwise the proxy doesn't get involved with file:// url's as they
are loaded directly off disk.

Assuming that part is handled, is there something that ensures the pages aren't
cached?

[nednguyen, 2016-09-30 16:46:20]

On 2016/09/30 16:39:52, Pat Meenan wrote:
> On 2016/09/30 09:50:45, nednguyen wrote:
> > On 2016/09/30 04:20:57, kouhei wrote:
> > > lgtm
> > 
> > Patrick: I suspect the ts_proxy's set command is not honored on windows:
> > 
> >   Traceback (most recent call last):
> >     File
> >
>
"C:\b\build\slave\catapult\build\catapult\telemetry\telemetry\page\page_run_end_to_end_unittest.py",
> > line 362, in testTrafficSettings
> >       latencies_by_page['fast'] + 5)
> >   AssertionError: 1.246000051498413 not greater than 5.3480000495910645
> > 
> > On my Mac & Linux station, using REGULAR_2G setting make it takes around 10s
> to
> > load the page. I would expect similar latency on windows.
> 
> I can reproduce the timings locally but I'm trying to understand exactly what
is
> being tested.  It looks like it is loading file://green_rect.html.  Is there
> some logic somewhere in telemetry that converts that into a http page to
> localhost?  Otherwise the proxy doesn't get involved with file:// url's as
they
> are loaded directly off disk.
> 
> Assuming that part is handled, is there something that ensures the pages
aren't
> cached?

The logic of the local mem cache server for serving the page is in
https://github.com/catapult-project/catapult/blob/master/telemetry/telemetry/...

When story_runner.Run(test, story_set, options, results) is invoked (line 359),
it creates the local server when encounter "file://" url here:
https://github.com/catapult-project/catapult/blob/master/telemetry/telemetry/...

[Pat Meenan, 2016-09-30 16:57:12]

On 2016/09/30 16:46:20, nednguyen wrote:
> On 2016/09/30 16:39:52, Pat Meenan wrote:
> > On 2016/09/30 09:50:45, nednguyen wrote:
> > > On 2016/09/30 04:20:57, kouhei wrote:
> > > > lgtm
> > > 
> > > Patrick: I suspect the ts_proxy's set command is not honored on windows:
> > > 
> > >   Traceback (most recent call last):
> > >     File
> > >
> >
>
"C:\b\build\slave\catapult\build\catapult\telemetry\telemetry\page\page_run_end_to_end_unittest.py",
> > > line 362, in testTrafficSettings
> > >       latencies_by_page['fast'] + 5)
> > >   AssertionError: 1.246000051498413 not greater than 5.3480000495910645
> > > 
> > > On my Mac & Linux station, using REGULAR_2G setting make it takes around
10s
> > to
> > > load the page. I would expect similar latency on windows.
> > 
> > I can reproduce the timings locally but I'm trying to understand exactly
what
> is
> > being tested.  It looks like it is loading file://green_rect.html.  Is there
> > some logic somewhere in telemetry that converts that into a http page to
> > localhost?  Otherwise the proxy doesn't get involved with file:// url's as
> they
> > are loaded directly off disk.
> > 
> > Assuming that part is handled, is there something that ensures the pages
> aren't
> > cached?
> 
> The logic of the local mem cache server for serving the page is in
>
https://github.com/catapult-project/catapult/blob/master/telemetry/telemetry/...
> 
> When story_runner.Run(test, story_set, options, results) is invoked (line
359),
> it creates the local server when encounter "file://" url here:
>
https://github.com/catapult-project/catapult/blob/master/telemetry/telemetry/...

Thanks.  One more quick question, why would you expect the page to take 5
seconds to load?  As best as I can tell, green_rect is a tiny, self-contained
html page and the 2G profile is setting the RTT to be 300ms.  Worst case I'd
expect to see it add ~600ms assuming a new connection is needed.

[nednguyen, 2016-09-30 17:16:50]

On 2016/09/30 16:57:12, Pat Meenan wrote:
> On 2016/09/30 16:46:20, nednguyen wrote:
> > On 2016/09/30 16:39:52, Pat Meenan wrote:
> > > On 2016/09/30 09:50:45, nednguyen wrote:
> > > > On 2016/09/30 04:20:57, kouhei wrote:
> > > > > lgtm
> > > > 
> > > > Patrick: I suspect the ts_proxy's set command is not honored on windows:
> > > > 
> > > >   Traceback (most recent call last):
> > > >     File
> > > >
> > >
> >
>
"C:\b\build\slave\catapult\build\catapult\telemetry\telemetry\page\page_run_end_to_end_unittest.py",
> > > > line 362, in testTrafficSettings
> > > >       latencies_by_page['fast'] + 5)
> > > >   AssertionError: 1.246000051498413 not greater than 5.3480000495910645
> > > > 
> > > > On my Mac & Linux station, using REGULAR_2G setting make it takes around
> 10s
> > > to
> > > > load the page. I would expect similar latency on windows.
> > > 
> > > I can reproduce the timings locally but I'm trying to understand exactly
> what
> > is
> > > being tested.  It looks like it is loading file://green_rect.html.  Is
there
> > > some logic somewhere in telemetry that converts that into a http page to
> > > localhost?  Otherwise the proxy doesn't get involved with file:// url's as
> > they
> > > are loaded directly off disk.
> > > 
> > > Assuming that part is handled, is there something that ensures the pages
> > aren't
> > > cached?
> > 
> > The logic of the local mem cache server for serving the page is in
> >
>
https://github.com/catapult-project/catapult/blob/master/telemetry/telemetry/...
> > 
> > When story_runner.Run(test, story_set, options, results) is invoked (line
> 359),
> > it creates the local server when encounter "file://" url here:
> >
>
https://github.com/catapult-project/catapult/blob/master/telemetry/telemetry/...
> 
> Thanks.  One more quick question, why would you expect the page to take 5
> seconds to load?  As best as I can tell, green_rect is a tiny, self-contained
> html page and the 2G profile is setting the RTT to be 300ms.  Worst case I'd
> expect to see it add ~600ms assuming a new connection is needed.

Oh, I think it takes about 5 seconds to load because I run it on my linux
station and it took about 10s :P

Now I think about it, that 10s may include the time it took to setup the
memcache server on Linux. If you think using the rtt 300ms is the best number, I
will change the test to use rtt(2G) - rtt(WiFi)

[nednguyen, 2016-09-30 18:51:31]

The CQ bit was checked by nednguyen@google.com to run a CQ dry run

[commit-bot: I haz the power, 2016-09-30 18:51:34]

Dry run: CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[nednguyen, 2016-09-30 19:05:21]

On 2016/09/30 17:16:50, nednguyen wrote:
> On 2016/09/30 16:57:12, Pat Meenan wrote:
> > On 2016/09/30 16:46:20, nednguyen wrote:
> > > On 2016/09/30 16:39:52, Pat Meenan wrote:
> > > > On 2016/09/30 09:50:45, nednguyen wrote:
> > > > > On 2016/09/30 04:20:57, kouhei wrote:
> > > > > > lgtm
> > > > > 
> > > > > Patrick: I suspect the ts_proxy's set command is not honored on
windows:
> > > > > 
> > > > >   Traceback (most recent call last):
> > > > >     File
> > > > >
> > > >
> > >
> >
>
"C:\b\build\slave\catapult\build\catapult\telemetry\telemetry\page\page_run_end_to_end_unittest.py",
> > > > > line 362, in testTrafficSettings
> > > > >       latencies_by_page['fast'] + 5)
> > > > >   AssertionError: 1.246000051498413 not greater than
5.3480000495910645
> > > > > 
> > > > > On my Mac & Linux station, using REGULAR_2G setting make it takes
around
> > 10s
> > > > to
> > > > > load the page. I would expect similar latency on windows.
> > > > 
> > > > I can reproduce the timings locally but I'm trying to understand exactly
> > what
> > > is
> > > > being tested.  It looks like it is loading file://green_rect.html.  Is
> there
> > > > some logic somewhere in telemetry that converts that into a http page to
> > > > localhost?  Otherwise the proxy doesn't get involved with file:// url's
as
> > > they
> > > > are loaded directly off disk.
> > > > 
> > > > Assuming that part is handled, is there something that ensures the pages
> > > aren't
> > > > cached?
> > > 
> > > The logic of the local mem cache server for serving the page is in
> > >
> >
>
https://github.com/catapult-project/catapult/blob/master/telemetry/telemetry/...
> > > 
> > > When story_runner.Run(test, story_set, options, results) is invoked (line
> > 359),
> > > it creates the local server when encounter "file://" url here:
> > >
> >
>
https://github.com/catapult-project/catapult/blob/master/telemetry/telemetry/...
> > 
> > Thanks.  One more quick question, why would you expect the page to take 5
> > seconds to load?  As best as I can tell, green_rect is a tiny,
self-contained
> > html page and the 2G profile is setting the RTT to be 300ms.  Worst case I'd
> > expect to see it add ~600ms assuming a new connection is needed.
> 
> Oh, I think it takes about 5 seconds to load because I run it on my linux
> station and it took about 10s :P
> 
> Now I think about it, that 10s may include the time it took to setup the
> memcache server on Linux. If you think using the rtt 300ms is the best number,
I
> will change the test to use rtt(2G) - rtt(WiFi)

PTAL again

[commit-bot: I haz the power, 2016-09-30 19:14:12]

The CQ bit was unchecked by commit-bot@chromium.org

[commit-bot: I haz the power, 2016-09-30 19:14:15]

Dry run: This issue passed the CQ dry run.

[Pat Meenan, 2016-09-30 20:59:00]

lgtm though the assert might be a little tight depending on how consistent the
timings are in general.  1ms of fluctuation would be enough to fail.

[nednguyen(REVIEW IN OTHER ACC), 2016-09-30 21:58:02]

The CQ bit was checked by nednguyen@chromium.org

[nednguyen(REVIEW IN OTHER ACC), 2016-09-30 21:58:02]

The patchset sent to the CQ was uploaded after l-g-t-m from kouhei@chromium.org
Link to the patchset: https://codereview.chromium.org/2377373002/#ps40001
(title: "Update test")

[commit-bot: I haz the power, 2016-09-30 21:58:12]

CQ is trying da patch. Follow status at
 
https://chromium-cq-status.appspot.com/v2/patch-status/codereview.chromium.or...

[commit-bot: I haz the power, 2016-09-30 21:59:48]

Description was changed from

==========
Add netconfigs for different network type

BUG=catapult:#2584
==========

to

==========
Add netconfigs for different network type

BUG=catapult:#2584

Committed:
https://chromium.googlesource.com/external/github.com/catapult-project/catapu...
==========

[commit-bot: I haz the power, 2016-09-30 21:59:49]

Committed patchset #3 (id:40001) as
https://chromium.googlesource.com/external/github.com/catapult-project/catapu...