Record accuracy of expected queueing time metric.

Record accuracy of expected queueing time metric.

We plan to use the expected queueing time metric to estimate renderer
main thread responsiveness from the renderer compositor. If the
renderer main thread is unresponsive, we'll then avoid blocking scroll
on input.

This patch measures how good the expected queueing time metric is
at predicting main thread responsiveness.

See the document here for details:
https://docs.google.com/document/d/1R4AMGyo1vyrfeCofqVSb06RqPExuRddVFDJdONenwt0/edit#heading=h.uamzit3mgd7h

BUG=599609
TEST=RenderWidgetHostLatencyTrackerTest.ExpectedQueueingTimeAccuracy

Review-Url: https://codereview.chromium.org/2954473002
Cr-Commit-Position: refs/heads/master@{#483169}
Committed: https://chromium.googlesource.com/chromium/src/+/f8d10a08eb9ae82564e45352f27bcff0197be110

[tdresser, 2017-06-22 12:08:14]

The CQ bit was checked by tdresser@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-06-22 12:08:21]

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, 2017-06-22 12:24:20]

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

[commit-bot: I haz the power, 2017-06-22 12:24:21]

Dry run: Try jobs failed on following builders:
  linux_chromium_tsan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[tdresser, 2017-06-22 14:26:28]

The CQ bit was checked by tdresser@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-06-22 14:26:44]

Dry run: CQ is trying da patch.

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

[tdresser, 2017-06-22 15:48:13]

tdresser@chromium.org changed reviewers:
+ aelias@chromium.org, altimin@chromium.org

[tdresser, 2017-06-22 15:48:14]

+aelias@ for general approach and content/.
+altimin@ for third_party/WebKit
+dtapuska@ FYI

[commit-bot: I haz the power, 2017-06-22 16:40:13]

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

[commit-bot: I haz the power, 2017-06-22 16:40:14]

Dry run: This issue passed the CQ dry run.

[aelias_OOO_until_Jul13, 2017-06-23 02:20:19]

lgtm

[tdresser, 2017-06-23 12:23:49]

tdresser@chromium.org changed reviewers:
+ jbauman@chromium.org, kenrb@chromium.org

[tdresser, 2017-06-23 12:23:50]

+jbauman@ for ui/latency.
+kenrb@ for ipc/mojo owners.

[kenrb, 2017-06-23 14:39:20]

mojo lgtm

[altimin, 2017-06-26 14:20:41]

lgtm

[tdresser, 2017-06-26 14:35:11]

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

[tdresser, 2017-06-26 14:35:12]

+isherman@ for histograms.xml

[Ilya Sherman, 2017-06-26 19:00:06]

This is an interesting metric design.  I think it's ok, but I'll also throw out
another idea: In the doc, you suggested measuring the ratio of actual to
expected queueing time.  You then noted that this might be noisy for low
queueing times.  Would it make sense, then, to measure the ratio for queueing
times above some threshold; and to measure the absolute difference for queueing
times below that threshold?  That seems like it might give you richer data, and
use fewer metrics to do so.  WDYT?

https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
File tools/metrics/histograms/histograms.xml (right):

https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
tools/metrics/histograms/histograms.xml:17171: +    Team:
input-dev@chromium.org.
Not sure whether you're moving this because of b/37249722; if so, note that that
bug is fixed :)

https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
tools/metrics/histograms/histograms.xml:60892: +   
name="RendererScheduler.QueueingDurationWhenExpectedQueueingTime.GreaterThan.10ms"
Please use histogram_suffixes to reduce the amount of repetition in this file.

[tdresser, 2017-06-27 13:45:15]

I thought some about using a ratio above some threshold, but I think for the
specific use case I'm looking at, using several cutoffs makes sense.

The plan
(https://docs.google.com/document/d/1_P6dw4YXePVXssUhg4DicU7T2f-6rt7A1OxniLvDt...)
is to use the Expected Queueing Time metric to decide when to intervene, and not
block input on the main thread. We're going to end up picking a threshold for
EQT, and this approach let's us look at the distribution of real queueing times
for the thresholds we're considering for the intervention.

https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
File tools/metrics/histograms/histograms.xml (right):

https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
tools/metrics/histograms/histograms.xml:17171: +    Team:
input-dev@chromium.org.
On 2017/06/26 19:00:06, Ilya Sherman wrote:
> Not sure whether you're moving this because of b/37249722; if so, note that
that
> bug is fixed :)

Ah. Well, I'll move it for consistency for now, and at some point might do a
bulk change moving all of these back as owners.

Sound reasonable?

https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
tools/metrics/histograms/histograms.xml:60892: +   
name="RendererScheduler.QueueingDurationWhenExpectedQueueingTime.GreaterThan.10ms"
On 2017/06/26 19:00:06, Ilya Sherman wrote:
> Please use histogram_suffixes to reduce the amount of repetition in this file.

Done.

[tdresser, 2017-06-27 13:45:28]

The CQ bit was checked by tdresser@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-06-27 13:45:34]

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, 2017-06-27 14:07:45]

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

[commit-bot: I haz the power, 2017-06-27 14:07:46]

Dry run: Try jobs failed on following builders:
  linux_chromium_compile_dbg_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[tdresser, 2017-06-27 14:14:34]

On 2017/06/27 13:45:15, tdresser wrote:
> I thought some about using a ratio above some threshold, but I think for the
> specific use case I'm looking at, using several cutoffs makes sense.
> 
> The plan
>
(https://docs.google.com/document/d/1_P6dw4YXePVXssUhg4DicU7T2f-6rt7A1OxniLvDt...)
> is to use the Expected Queueing Time metric to decide when to intervene, and
not
> block input on the main thread. We're going to end up picking a threshold for
> EQT, and this approach let's us look at the distribution of real queueing
times
> for the thresholds we're considering for the intervention.
> 
>
https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
> File tools/metrics/histograms/histograms.xml (right):
> 
>
https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
> tools/metrics/histograms/histograms.xml:17171: +    Team:
> mailto:input-dev@chromium.org.
> On 2017/06/26 19:00:06, Ilya Sherman wrote:
> > Not sure whether you're moving this because of b/37249722; if so, note that
> that
> > bug is fixed :)
> 
> Ah. Well, I'll move it for consistency for now, and at some point might do a
> bulk change moving all of these back as owners.
> 
> Sound reasonable?
> 
>
https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
> tools/metrics/histograms/histograms.xml:60892: +   
>
name="RendererScheduler.QueueingDurationWhenExpectedQueueingTime.GreaterThan.10ms"
> On 2017/06/26 19:00:06, Ilya Sherman wrote:
> > Please use histogram_suffixes to reduce the amount of repetition in this
file.
> 
> Done.

Sorry most recent patchset is messed up, hold off on re-review for a bit.

[tdresser, 2017-06-27 14:18:32]

Patchset #3 (id:40001) has been deleted

[tdresser, 2017-06-27 14:21:20]

The CQ bit was checked by tdresser@chromium.org to run a CQ dry run

[commit-bot: I haz the power, 2017-06-27 14:21:27]

Dry run: CQ is trying da patch.

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

[tdresser, 2017-06-27 14:21:44]

On 2017/06/27 14:14:34, tdresser wrote:
> On 2017/06/27 13:45:15, tdresser wrote:
> > I thought some about using a ratio above some threshold, but I think for the
> > specific use case I'm looking at, using several cutoffs makes sense.
> > 
> > The plan
> >
>
(https://docs.google.com/document/d/1_P6dw4YXePVXssUhg4DicU7T2f-6rt7A1OxniLvDt...)
> > is to use the Expected Queueing Time metric to decide when to intervene, and
> not
> > block input on the main thread. We're going to end up picking a threshold
for
> > EQT, and this approach let's us look at the distribution of real queueing
> times
> > for the thresholds we're considering for the intervention.
> > 
> >
>
https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
> > File tools/metrics/histograms/histograms.xml (right):
> > 
> >
>
https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
> > tools/metrics/histograms/histograms.xml:17171: +    Team:
> > mailto:input-dev@chromium.org.
> > On 2017/06/26 19:00:06, Ilya Sherman wrote:
> > > Not sure whether you're moving this because of b/37249722; if so, note
that
> > that
> > > bug is fixed :)
> > 
> > Ah. Well, I'll move it for consistency for now, and at some point might do a
> > bulk change moving all of these back as owners.
> > 
> > Sound reasonable?
> > 
> >
>
https://codereview.chromium.org/2954473002/diff/20001/tools/metrics/histogram...
> > tools/metrics/histograms/histograms.xml:60892: +   
> >
>
name="RendererScheduler.QueueingDurationWhenExpectedQueueingTime.GreaterThan.10ms"
> > On 2017/06/26 19:00:06, Ilya Sherman wrote:
> > > Please use histogram_suffixes to reduce the amount of repetition in this
> file.
> > 
> > Done.
> 
> Sorry most recent patchset is messed up, hold off on re-review for a bit.

It's good to go now.

[Ilya Sherman, 2017-06-27 15:12:33]

Metrics LGTM, thanks.

[commit-bot: I haz the power, 2017-06-27 16:30:36]

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

[commit-bot: I haz the power, 2017-06-27 16:30:37]

Dry run: This issue passed the CQ dry run.

[jbauman, 2017-06-27 21:59:10]

lgtm

[tdresser, 2017-06-28 12:23:31]

The CQ bit was checked by tdresser@chromium.org

[tdresser, 2017-06-28 12:23:32]

The patchset sent to the CQ was uploaded after l-g-t-m from kenrb@chromium.org,
aelias@chromium.org, altimin@chromium.org
Link to the patchset: https://codereview.chromium.org/2954473002/#ps60001
(title: "Use histogram suffixes, add missing test coverage.")

[commit-bot: I haz the power, 2017-06-28 12:23:45]

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, 2017-06-28 12:30:47]

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

[commit-bot: I haz the power, 2017-06-28 12:30:48]

Try jobs failed on following builders:
  chromium_presubmit on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[tdresser, 2017-06-28 13:54:08]

The CQ bit was checked by tdresser@chromium.org

[tdresser, 2017-06-28 13:54:09]

The patchset sent to the CQ was uploaded after l-g-t-m from kenrb@chromium.org,
aelias@chromium.org, isherman@chromium.org, altimin@chromium.org,
jbauman@chromium.org
Link to the patchset: https://codereview.chromium.org/2954473002/#ps80001
(title: "Rebase.")

[commit-bot: I haz the power, 2017-06-28 13:54:25]

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, 2017-06-28 14:03:26]

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

[commit-bot: I haz the power, 2017-06-28 14:03:27]

Try jobs failed on following builders:
  chromium_presubmit on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/chromium_presub...)

[tdresser, 2017-06-28 15:36:05]

tdresser@chromium.org changed reviewers:
+ rockot@chromium.org

[tdresser, 2017-06-28 15:36:06]

+rockot for added dependency on mojo/common.

[Ken Rockot(use gerrit already), 2017-06-28 16:01:12]

lgtm

[tdresser, 2017-06-28 16:55:15]

The CQ bit was checked by tdresser@chromium.org

[commit-bot: I haz the power, 2017-06-28 16:55:37]

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, 2017-06-28 17:44:14]

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

[commit-bot: I haz the power, 2017-06-28 17:44:16]

Try jobs failed on following builders:
  linux_chromium_tsan_rel_ng on master.tryserver.chromium.linux (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)

[tdresser, 2017-06-28 18:51:34]

The CQ bit was checked by tdresser@chromium.org

[tdresser, 2017-06-28 18:51:35]

The patchset sent to the CQ was uploaded after l-g-t-m from rockot@chromium.org,
aelias@chromium.org, altimin@chromium.org, kenrb@chromium.org,
jbauman@chromium.org, isherman@chromium.org
Link to the patchset: https://codereview.chromium.org/2954473002/#ps100001
(title: "Set trace ID in test.")

[commit-bot: I haz the power, 2017-06-28 18:52:04]

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, 2017-06-28 19:42:24]

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

[commit-bot: I haz the power, 2017-06-28 19:42:26]

Try jobs failed on following builders:
  linux_chromium_chromeos_ozone_rel_ng on master.tryserver.chromium.linux
(JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.linux/builders/linux_chromium_...)
  win_chromium_rel_ng on master.tryserver.chromium.win (JOB_FAILED,
http://build.chromium.org/p/tryserver.chromium.win/builders/win_chromium_rel_...)

[tdresser, 2017-06-28 19:44:04]

The CQ bit was checked by tdresser@chromium.org

[commit-bot: I haz the power, 2017-06-28 19:44:31]

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, 2017-06-28 22:18:04]

CQ is committing da patch.

Bot data: {"patchset_id": 100001, "attempt_start_ts": 1498679044678310,
"parent_rev": "fd543d7dd8eb14641e5f48ce7e23d3f20499a73b", "commit_rev":
"0976ed0c8cb86ac457ba3408554dc0ae0a42fbee"}

[commit-bot: I haz the power, 2017-06-28 22:18:28]

CQ is committing da patch.

Bot data: {"patchset_id": 100001, "attempt_start_ts": 1498679044678310,
"parent_rev": "d0d190c09619cb359296999438551b66f0e1cdb1", "commit_rev":
"f8d10a08eb9ae82564e45352f27bcff0197be110"}

[commit-bot: I haz the power, 2017-06-28 22:18:42]

Description was changed from

==========
Record accuracy of expected queueing time metric.

We plan to use the expected queueing time metric to estimate renderer
main thread responsiveness from the renderer compositor. If the
renderer main thread is unresponsive, we'll then avoid blocking scroll
on input.

This patch measures how good the expected queueing time metric is
at predicting main thread responsiveness.

See the document here for details:
https://docs.google.com/document/d/1R4AMGyo1vyrfeCofqVSb06RqPExuRddVFDJdONenw...

BUG=599609
TEST=RenderWidgetHostLatencyTrackerTest.ExpectedQueueingTimeAccuracy
==========

to

==========
Record accuracy of expected queueing time metric.

We plan to use the expected queueing time metric to estimate renderer
main thread responsiveness from the renderer compositor. If the
renderer main thread is unresponsive, we'll then avoid blocking scroll
on input.

This patch measures how good the expected queueing time metric is
at predicting main thread responsiveness.

See the document here for details:
https://docs.google.com/document/d/1R4AMGyo1vyrfeCofqVSb06RqPExuRddVFDJdONenw...

BUG=599609
TEST=RenderWidgetHostLatencyTrackerTest.ExpectedQueueingTimeAccuracy

Review-Url: https://codereview.chromium.org/2954473002
Cr-Commit-Position: refs/heads/master@{#483169}
Committed:
https://chromium.googlesource.com/chromium/src/+/f8d10a08eb9ae82564e45352f27b...
==========

[commit-bot: I haz the power, 2017-06-28 22:18:46]

Committed patchset #5 (id:100001) as
https://chromium.googlesource.com/chromium/src/+/f8d10a08eb9ae82564e45352f27b...