Looking at logcat output from one of the failures:

"""
https://luci-logdog.appspot.com/logs/chromium/android/swarming/logcats/41a3ca2dff6f6811/+/logcat_logcat_org.chromium.chrome.browser.vr.WebXrVrInputTest.testScreenTapsRegisteredOnCardboard_WebXr__ChromeTabbedActivity_20181207T234321-UTC_01e19439caf03c4d
"""

nothing obviously chrome related.

billorr@, klausw@: could you investigate these failures? They're impacting the CQ, so marking P1.

Taking a look, but at least recently, the only failures I'm seeing are on a specific patch (implying it's that patch's fault), and I'm unable to load the tombstones from the linked failures. I'll keep digging.

Logcat doesn't have anything particularly useful, just that Chrome is dying without providing a stack trace.

Do you know how long this has been going on for? I upgraded the VR test APKs, which might potentially affect this (although logcat says we're successfully loading VrCore), but those upgrades went in > 1 week ago.

> Do you know how long this has been going on for?

I have not investigated. Try reaching out to CCI team [+bradhall], although they may not easily know the answer either.

I have a patch out to disable the failing tests while I continue to investigate. Not much to go on ATM other than the fact that all the failing tests are WebXR-related (as opposed to WebVR).

Looks like whatever's happening is affecting more than just Chrome?

We get the following line when Chrome crashes: Process org.chromium.chrome:privileged_process0 (pid 24863) has died

But we get the same line for a number of other apps in the same logcat:
com.google.android.talk
com.google.android.youtube
com.qualcomm.qcrilmsgtunnel
com.google.android.gms.persistent
com.google.android.inputmethod.latin
com.google.android.gms.wearable
com.google.android.googlequicksearchbox:search
com.google.process.gapps

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/4a2f6274a5b91d1bcba93db4ee5756c6e6de5255

commit 4a2f6274a5b91d1bcba93db4ee5756c6e6de5255
Author: bsheedy <bsheedy@chromium.org>
Date: Tue Dec 11 21:34:41 2018

Disable failing WebXR tests

Disables failing WebXR tests that are causing CQ issues. Will be
re-enabled once the root cause is found and fixed.

TBR=billorr@chromium.org

Bug:  913161 
Change-Id: If50af30cc9e24ce917989a0b0ac210254ce503bb
Reviewed-on: https://chromium-review.googlesource.com/c/1372540
Reviewed-by: Brian Sheedy <bsheedy@chromium.org>
Commit-Queue: Brian Sheedy <bsheedy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#615670}
[modify] https://crrev.com/4a2f6274a5b91d1bcba93db4ee5756c6e6de5255/chrome/android/javatests/src/org/chromium/chrome/browser/vr/WebXrVrInputTest.java
[modify] https://crrev.com/4a2f6274a5b91d1bcba93db4ee5756c6e6de5255/chrome/android/javatests/src/org/chromium/chrome/browser/vr/WebXrVrTransitionTest.java

billorr@ pointed out that this appears to be due to running out of memory, as there are log messages from lowmemorykiller. Still not sure why this only happens with WebXR tests, though.

Bisect points to https://chromium-review.googlesource.com/c/1364164 as the culprit.

ericrk@, any idea why this would be causing what looks like memory issues while in certain VR situations?

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/1caddf18637d3834d255fdd52d968de038c7b944

commit 1caddf18637d3834d255fdd52d968de038c7b944
Author: Eric Karl <ericrk@chromium.org>
Date: Fri Dec 14 18:42:30 2018

Revert "Disable failing WebXR tests"

This reverts commit 4a2f6274a5b91d1bcba93db4ee5756c6e6de5255.

Reason for revert: Reverted the CL causing failures, so re-enabling tests.

Original change's description:
> Disable failing WebXR tests
> 
> Disables failing WebXR tests that are causing CQ issues. Will be
> re-enabled once the root cause is found and fixed.
> 
> TBR=billorr@chromium.org
> 
> Bug:  913161 
> Change-Id: If50af30cc9e24ce917989a0b0ac210254ce503bb
> Reviewed-on: https://chromium-review.googlesource.com/c/1372540
> Reviewed-by: Brian Sheedy <bsheedy@chromium.org>
> Commit-Queue: Brian Sheedy <bsheedy@chromium.org>
> Cr-Commit-Position: refs/heads/master@{#615670}

TBR=bsheedy@chromium.org,billorr@chromium.org

# Not skipping CQ checks because original CL landed > 1 day ago.

Bug:  913161 
Change-Id: I309819ee2a135d8ec30cc424726daf4a4fa15a2f
Reviewed-on: https://chromium-review.googlesource.com/c/1377489
Reviewed-by: Bill Orr <billorr@chromium.org>
Reviewed-by: Brian Sheedy <bsheedy@chromium.org>
Reviewed-by: Eric Karl <ericrk@chromium.org>
Commit-Queue: Brian Sheedy <bsheedy@chromium.org>
Cr-Commit-Position: refs/heads/master@{#616760}
[modify] https://crrev.com/1caddf18637d3834d255fdd52d968de038c7b944/chrome/android/javatests/src/org/chromium/chrome/browser/vr/WebXrVrInputTest.java
[modify] https://crrev.com/1caddf18637d3834d255fdd52d968de038c7b944/chrome/android/javatests/src/org/chromium/chrome/browser/vr/WebXrVrTransitionTest.java

Tests are re-enabled and should be passing consistently again.

I'll work with ericrk@ to repro/diagnose the issue with the culprit CL.

An alternate fix has landed for the last two weeks and these tests do not seem to be impacted any more. Closing this out.