All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=592299

Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?keys=agxzfmNocm9tZXBlcmZyFAsSB0Fub21hbHkYgIDA36_BvwoM


Bot(s) for this bug's original alert(s):

linux-release

=== Auto-CCing suspected CL author brucedawson@chromium.org ===

Hi brucedawson@chromium.org, the bisect results pointed to your CL below as possibly
causing a regression. Please have a look at this info and see whether
your CL be related.


===== BISECT JOB RESULTS =====
Status: completed


===== SUSPECTED CL(s) =====
Subject : Roll src\third_party\leveldatabase\src/ 706b7f8d4..7306ef856 (11 commits).
Author  : brucedawson
Commit description:
  
https://chromium.googlesource.com/external/leveldb.git/+log/706b7f8d43b0..7306ef856a91

    $ git log 706b7f8d4..7306ef856 --date=short --no-merges --format='%ad %ae %s'
    2016-02-19 brucedawson Fix signed/unsigned mismatch on VC++ builds
    2016-01-29 cmumford Putting build artifacts in subdirectory.
    2016-01-15 ralphtheninja add travis build badge
    2016-01-14 cmumford Added a Travis CI build file.
    2016-01-04 cmumford Added a contributors section to README.md
    2015-11-23 venilton.falvo Update README.md
    2015-04-20 ideawu fix indent
    2015-04-20 ideawu log compaction output file's level along with number
    2015-02-17 paul.irish documentation. improved link
    2015-01-10 paul.irish readme: improved documentation link
    2014-12-17 digit Fix Android/MIPS build.

BUG= 582658 

Review URL: https://codereview.chromium.org/1740533002

Cr-Commit-Position: refs/heads/master@{#379170}
Commit  : 588e9b5d83f861fd5ab21dad7746fd0e74184999
Date    : Fri Mar 04 01:43:40 2016


===== TESTED REVISIONS =====
Revision                Mean Value  Std. Dev.   Num Values  Good?
chromium@379161         85.789744   2.53427     39          good
chromium@379167         87.025      1.736787    8           good
chromium@379169         86.2375     1.51934     8           good
chromium@379170         84.06       0.31305     5           bad
chromium@379173         84.307143   1.178959    14          bad
chromium@379184         84.6525     1.960375    40          bad

Bisect job ran on: linux_perf_bisect
Bug ID: 592299

Test Command: src/tools/perf/run_benchmark -v --browser=release --output-format=chartjson --also-run-disabled-tests blink_perf.pywebsocket
Test Metric: fetch-send-text-window-async-verify/fetch-send-text-window-async-verify
Relative Change: 0.19%
Score: 99.5

Buildbot stdio: http://build.chromium.org/p/tryserver.chromium.perf/builders/linux_perf_bisect/builds/6284
Job details: https://chromeperf.appspot.com/buildbucket_job_status/9018878279620182464


| O O | Visit http://www.chromium.org/developers/speed-infra/perf-bug-faq
|  X  | for more information addressing perf regression bugs. For feedback,
| / \ | file a bug with label Cr-Tests-AutoBisect.  Thank you!

Sorry Bruce, looks like the bisect found a 0.19% change, it was supposed to be looking for a 20% change. We're working on tuning the bisect's algorithm so that it stops cc-ing authors of small changes.

Assigning to tyoshino, owner of blink_perf.pywebsocket benchmark. Looking at the graph on linux, it appears to be bimodal, occasionally switching from usually on the high end to usually on the low end. On this last switch, we weren't able to get any culprit CL in the bisect. It's possible there isn't a regression, just a problem with the test. Any ideas?

Note that it would be easier to understand what's going on if someone could look into enabling the ref build ( bug 530374 )

tyoshino@, gentle ping..

If it looks like there are any regressions from this CL them cmumford@ should be brought in. The roll pulled in quite a few changes but only one was actually mine and they might have more context for the other changes.

The land/revert pattern follows f68b7c07a30baac088b8eda055ba52835c04692c and b3521bc11841b49542d4b4c6a415167202ae151b. atotic, just FYI for you that your original patch had this positive impact.