Log a non-crash dump for 10% of Spitzer pipeline decode failures.

Log a non-crash dump for 10% of Spitzer pipeline decode failures.

Spitzer is currently enabled as a 40% experiment on dev (soon on
beta as well), but we have abnormally high video-only decode
failure rates, so add some logging to try and track this down.

~10% of the time, this change will log the 'subresource_url' crash
key as a non-crash dump when a decode error is encountered on
Android where Spitzer is enabled.

This will allow us to investigate and triage these crashes. We're
already logging this key for normal crash dumps during DoLoad().

This is a temporary change which will be removed shortly after
landing once any issues are identified.

BUG=595076
TEST=none

Committed: https://crrev.com/6e3901e3f6240d8660cb039fdcbb1dd6f75121f6
Cr-Commit-Position: refs/heads/master@{#381532}

[DaleCurtis, 2016-03-15 19:44:08]

Description was changed from

==========
Log a non-crash dump for 10% of Android pipeline decode failures.

This will allow us to investigate and triage these crashes. We're
already logging this key for normal crash dumps during DoLoad().

This is a temporary change which will be removed shortly after
landing once any issues are identified.

BUG=595076
TEST=none
==========

to

==========
Log a non-crash dump for 10% of Spitzer pipeline decode failures.

Spitzer is currently enabled as a 40% experiment on dev (soon on
beta as well), but we have abnormally high video-only decode
failure rates, so add some logging to try and track this down.

~10% of the time, this change will log the 'subresource_url' crash
key as a non-crash dump when a decode error is encountered on
Android where Spitzer is enabled.

This will allow us to investigate and triage these crashes. We're
already logging this key for normal crash dumps during DoLoad().

This is a temporary change which will be removed shortly after
landing once any issues are identified.

BUG=595076
TEST=none
==========

[DaleCurtis, 2016-03-15 19:45:50]

dalecurtis@chromium.org changed reviewers:
+ rsesek@chromium.org

[DaleCurtis, 2016-03-15 19:45:51]

+rsesek for review since this involves crash keys.

[Robert Sesek, 2016-03-16 18:11:01]

lgtm

[DaleCurtis, 2016-03-16 18:11:51]

The CQ bit was checked by dalecurtis@chromium.org

[commit-bot: I haz the power, 2016-03-16 18:12:23]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1808473002/1
View timeline at
 https://chromium-cq-status.appspot.com/patch-timeline/1808473002/1

[commit-bot: I haz the power, 2016-03-16 20:51:22]

Committed patchset #1 (id:1)

[commit-bot: I haz the power, 2016-03-16 20:53:25]

Description was changed from

==========
Log a non-crash dump for 10% of Spitzer pipeline decode failures.

Spitzer is currently enabled as a 40% experiment on dev (soon on
beta as well), but we have abnormally high video-only decode
failure rates, so add some logging to try and track this down.

~10% of the time, this change will log the 'subresource_url' crash
key as a non-crash dump when a decode error is encountered on
Android where Spitzer is enabled.

This will allow us to investigate and triage these crashes. We're
already logging this key for normal crash dumps during DoLoad().

This is a temporary change which will be removed shortly after
landing once any issues are identified.

BUG=595076
TEST=none
==========

to

==========
Log a non-crash dump for 10% of Spitzer pipeline decode failures.

Spitzer is currently enabled as a 40% experiment on dev (soon on
beta as well), but we have abnormally high video-only decode
failure rates, so add some logging to try and track this down.

~10% of the time, this change will log the 'subresource_url' crash
key as a non-crash dump when a decode error is encountered on
Android where Spitzer is enabled.

This will allow us to investigate and triage these crashes. We're
already logging this key for normal crash dumps during DoLoad().

This is a temporary change which will be removed shortly after
landing once any issues are identified.

BUG=595076
TEST=none

Committed: https://crrev.com/6e3901e3f6240d8660cb039fdcbb1dd6f75121f6
Cr-Commit-Position: refs/heads/master@{#381532}
==========

[commit-bot: I haz the power, 2016-03-16 20:53:26]

Patchset 1 (id:??) landed as
https://crrev.com/6e3901e3f6240d8660cb039fdcbb1dd6f75121f6
Cr-Commit-Position: refs/heads/master@{#381532}

[DaleCurtis, 2016-03-29 02:09:32]

rsesek: Are non-crash dumps working on Android? This has not fired since it
landed which is pretty impossible to believe given our UMA metrics which show
plenty of decode errors on Android.

[Robert Sesek, 2016-03-29 20:26:53]

On 2016/03/29 02:09:32, DaleCurtis wrote:
> rsesek: Are non-crash dumps working on Android? This has not fired since it
> landed which is pretty impossible to believe given our UMA metrics which show
> plenty of decode errors on Android.

As far as I'm aware, it should be. Try a local build and see if it generates a
dump?

[DaleCurtis, 2016-04-12 23:42:15]

A revert of this CL (patchset #1 id:1) has been created in
https://codereview.chromium.org/1885733002/ by dalecurtis@chromium.org.

The reason for reverting is: Should have been reverted before M-51 branch, but
forgot..

[DaleCurtis, 2016-04-12 23:45:24]

rsesek@ I don't think dump-without-crashing works in Android renderer processes:

https://crash.corp.google.com/browse?q=product.name%3D%27Chrome_Android%27%20...

There are no reports since Chrome 42 || 44 for renderer-initiated dumps.

[Robert Sesek, 2016-04-13 14:46:21]

On 2016/04/12 23:45:24, DaleCurtis wrote:
> rsesek@ I don't think dump-without-crashing works in Android renderer
processes:
> 
>
https://crash.corp.google.com/browse?q=product.name%3D%27Chrome_Android%27%20...
> 
> There are no reports since Chrome 42 || 44 for renderer-initiated dumps.

That's a bag of sad. Can you file a bug for that and we can maybe get someone to
take a look (I've got no bandwidth right now).

[DaleCurtis, 2016-04-13 20:56:13]

Done, https://bugs.chromium.org/p/chromium/issues/detail?id=603272