UMA histogram for longest renderer main thread usage for page save to MHTML.

UMA histogram for longest renderer main thread usage for page save to MHTML.

The information about the time taken by the main render process thread
for saving a page to MHTML is the main feedback we have to see how much
the operation impacts user experience.

The existing UMA that tracks this information,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SingleFrame,
is not directly useful for the purpose above because it includes reports
for all frames in a page save operation. For pages with many frames, most
of them are very simple and quickly saved, while one -- generally the
main frame -- is the slowest. So we have a huge amount of low-value
samples pulling the statistics down.

This change adds a new histogram,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SlowestFrame,
that reports only the longest of all times from the former one in a
single HTML page saving operation, which is the most important one in
terms of user experience.

BUG=645686
Committed: https://crrev.com/90bf4f0c77709ddb65b1f8d86e04065aacae02ac
Cr-Commit-Position: refs/heads/master@{#424312}

[carlosk, 2016-10-08 00:50:07]

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

[commit-bot: I haz the power, 2016-10-08 00:50:46]

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

[carlosk, 2016-10-08 00:53:06]

carlosk@chromium.org changed reviewers:
+ isherman@chromium.org, lukasza@chromium.org, nasko@chromium.org

[carlosk, 2016-10-08 00:53:06]

isherman@: Please review changes in histograms.xml.

lukasza@, nasko@: Please review changes in mhtml_generation_manager.cc.

[carlosk, 2016-10-08 00:54:55]

Description was changed from

==========
UMA histogram for longest renderer main thread usage for page save to MHTML.

The information about the time taken by the main render process thread
for saving a page to MHTML is the main feedback we have to see how much
the operation impacts user experience.

The existing UMA that tracks this information,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SingleFrame,
is not directly useful for the purpose above because it includes reports
for all frames in a page save operation. For pages with many frames, most
of them are very simple and quickly saved, while one -- generally the
main frame -- is the slowest. Se we have a huge amount of low-value
samples pulling the statistics down.

This change adds a new histogram,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SlowestFrame,
that reports only the longest of all times from the former one in a
single HTML page saving operation, which is the most important one in
terms of user experience.

BUG=645686
==========

to

==========
UMA histogram for longest renderer main thread usage for page save to MHTML.

The information about the time taken by the main render process thread
for saving a page to MHTML is the main feedback we have to see how much
the operation impacts user experience.

The existing UMA that tracks this information,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SingleFrame,
is not directly useful for the purpose above because it includes reports
for all frames in a page save operation. For pages with many frames, most
of them are very simple and quickly saved, while one -- generally the
main frame -- is the slowest. So we have a huge amount of low-value
samples pulling the statistics down.

This change adds a new histogram,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SlowestFrame,
that reports only the longest of all times from the former one in a
single HTML page saving operation, which is the most important one in
terms of user experience.

BUG=645686
==========

[Łukasz Anforowicz, 2016-10-08 01:38:40]

LGTM

[commit-bot: I haz the power, 2016-10-08 01:56:26]

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

[commit-bot: I haz the power, 2016-10-08 01:56:27]

Dry run: This issue passed the CQ dry run.

[nasko, 2016-10-10 22:10:31]

Rubberstamp LGTM based on lukasza@'s review.

[Ilya Sherman, 2016-10-10 23:32:47]

lgtm

[Ilya Sherman, 2016-10-10 23:32:54]

The CQ bit was checked by isherman@chromium.org

[commit-bot: I haz the power, 2016-10-10 23:33:13]

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, 2016-10-11 00:50:23]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-10-11 00:54:25]

Description was changed from

==========
UMA histogram for longest renderer main thread usage for page save to MHTML.

The information about the time taken by the main render process thread
for saving a page to MHTML is the main feedback we have to see how much
the operation impacts user experience.

The existing UMA that tracks this information,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SingleFrame,
is not directly useful for the purpose above because it includes reports
for all frames in a page save operation. For pages with many frames, most
of them are very simple and quickly saved, while one -- generally the
main frame -- is the slowest. So we have a huge amount of low-value
samples pulling the statistics down.

This change adds a new histogram,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SlowestFrame,
that reports only the longest of all times from the former one in a
single HTML page saving operation, which is the most important one in
terms of user experience.

BUG=645686
==========

to

==========
UMA histogram for longest renderer main thread usage for page save to MHTML.

The information about the time taken by the main render process thread
for saving a page to MHTML is the main feedback we have to see how much
the operation impacts user experience.

The existing UMA that tracks this information,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SingleFrame,
is not directly useful for the purpose above because it includes reports
for all frames in a page save operation. For pages with many frames, most
of them are very simple and quickly saved, while one -- generally the
main frame -- is the slowest. So we have a huge amount of low-value
samples pulling the statistics down.

This change adds a new histogram,
PageSerialization.MhtmlGeneration.RendererMainThreadTime.SlowestFrame,
that reports only the longest of all times from the former one in a
single HTML page saving operation, which is the most important one in
terms of user experience.

BUG=645686

Committed: https://crrev.com/90bf4f0c77709ddb65b1f8d86e04065aacae02ac
Cr-Commit-Position: refs/heads/master@{#424312}
==========

[commit-bot: I haz the power, 2016-10-11 00:54:26]

Patchset 1 (id:??) landed as
https://crrev.com/90bf4f0c77709ddb65b1f8d86e04065aacae02ac
Cr-Commit-Position: refs/heads/master@{#424312}