--> nednguyen, for perf waterfall

I'm not sure what happened here. The timing doesn't seem to correspond with the OBBS migration, so probably not related. Looking at the first failing run:
https://ci.chromium.org/buildbot/chromium.perf/Win%207%20Nvidia%20GPU%20Perf/2509

Looks like the trigger step is there, but not the step where it collects and uploads the results.

dtu/ned,

What was the timing of the migration? Just wondering. The bots all broke sometime right after the last OK build so maybe the upload broke at a point when the whole build was broken.

jmadill@ the migration happen around July 09th with https://chromium-review.googlesource.com/c/chromium/src/+/1129364

Hmhh, the log says there is no problem with uploading the data to perf dashboard: https://logs.chromium.org/v/?s=chrome%2Fbb%2Fchromium.perf%2FWin_7_Nvidia_GPU_Perf%2F2940%2F%2B%2Frecipes%2Fsteps%2Fpassthrough_command_buffer_perftests_on_NVIDIA_GPU_on_Windows_on_Windows-2008ServerR2-SP1%2F0%2Flogs%2FMerge_script_log%2F0

Ah, look like this benchmark is somehow renamed to command_buffer_perftests

https://chromeperf.appspot.com/report?sid=c8d131d70fa7cfa5c2351f804020a00e608d5028e1f7480a1b8bbcade7dc95b1

Ned: I see "Duration of command_buffer_perftests upload time: 4 seconds"

Shouldn't that be passthrough_command_buffer_perftests or validating_command_buffer_perftests? Maybe that's the problem.

Yeah, that seems to be the issue. Lemme look at it

Ok, so the problem is the test step is named "passthrough_command_buffer_perftests", but the test command is still command_buffer_perftests. Hence, the data are produced in a folder named command_buffer_perftests/, which make the script uploading to perf dashboard think this benchmark's name is command_buffer_perftests

This is the line that's problematic: https://cs.chromium.org/chromium/src/testing/scripts/run_performance_tests.py?rcl=e7e06bb0214dd2c884dfc6f041683cc84e96a869&l=197

Nice work getting to the bottom of this Ned. Looking forward to the fix.

The following revision refers to this bug:
  https://chromium.googlesource.com/chromium/src.git/+/3a1d09a18914c7e4f61d3392291136eeadb9d7dd

commit 3a1d09a18914c7e4f61d3392291136eeadb9d7dd
Author: Ned Nguyen <nednguyen@google.com>
Date: Sat Aug 04 01:10:51 2018

Ensure gtest benchmark's name is the same as its corresponding test step's name

Bug:870692,757933
Cq-Include-Trybots: master.tryserver.chromium.perf:obbs_fyi
Change-Id: Ifc70a71cf4ddb2628038b45b51a57381a2a41c3a

NOTRY=true  # CQ flake

Change-Id: Ifc70a71cf4ddb2628038b45b51a57381a2a41c3a
Reviewed-on: https://chromium-review.googlesource.com/1162584
Commit-Queue: Ned Nguyen <nednguyen@google.com>
Reviewed-by: David Tu <dtu@chromium.org>
Reviewed-by: John Budorick <jbudorick@chromium.org>
Cr-Commit-Position: refs/heads/master@{#580724}
[modify] https://crrev.com/3a1d09a18914c7e4f61d3392291136eeadb9d7dd/testing/buildbot/chromium.perf.fyi.json
[modify] https://crrev.com/3a1d09a18914c7e4f61d3392291136eeadb9d7dd/testing/buildbot/chromium.perf.json
[modify] https://crrev.com/3a1d09a18914c7e4f61d3392291136eeadb9d7dd/testing/scripts/run_performance_tests.py
[modify] https://crrev.com/3a1d09a18914c7e4f61d3392291136eeadb9d7dd/tools/perf/core/perf_data_generator.py
[modify] https://crrev.com/3a1d09a18914c7e4f61d3392291136eeadb9d7dd/tools/perf/core/perf_data_generator_unittest.py

jmadill@: the graph to look for is https://chromeperf.appspot.com/report?sid=125205e486d1caa466a6fdfc73eba876bea0e16919e99d89732d3109d1558de6 

Thanks for the fix and pointer to the new reports.