Checking closer, the release builder failures tend to be on x86 platforms, but we're also seeing less common failures (such as issue 694344) on the kevin release builder / kevin paladin.

Killing numerous CQ runs left right and center, raising to P0

Recent example https://luci-milo.appspot.com/buildbot/chromeos/kevin-paladin/271

This is a bug in the test or in the product, so we want a Sheriff
to deal with it.  Last I knew, shawnn@ was looking at moving the
test to bvt-perbuild.  We should probably move that forward, unless
the test owners have a fix we can apply sooner.

failed on https://luci-milo.appspot.com/buildbot/chromeos/veyron_speedy-paladin/4459 too

The following revision refers to this bug:
  https://chrome-internal.googlesource.com/chromeos/autotest-cheets/+/15dc159fc4fabf3427b65a1de0d2eefca3738b45

commit 15dc159fc4fabf3427b65a1de0d2eefca3738b45
Author: Shawn Nematbakhsh <shawnn@chromium.org>
Date: Wed Feb 22 20:31:19 2017

Re-assign for triage.

https://wmatrix.googleplex.com/unfiltered?hide_missing=True&tests=cheets_MediaPlayerVideoHWDecodeUsed&days_back=30

It becomes flaky since R58-9294.0.0. I think it's just another victim of the issue https://bugs.chromium.org/p/chromium/issues/detail?id=694094.

The fix is in chrome, and was merged already. But chrome is not uprev for almost a week.

https://cros-goldeneye.corp.google.com/chromeos/console/listBuild#%2F

That's why we kept seeing this test failed.

Adding this weeks gardeners. Perhaps a manual uprev is necessary?

+derat who is thinking a lot about cq vs pfq test coverage

While failures on 58 should be due to  issue 694094 , we seem to have another group of failures in this test on 57 as well, which should be a different issue.

It's a dup to crbug.com/683053.
The fix CL is merged few hours ago in M57.