Understand how reshard perf benchmarks create regressions that are bisectable by pinpoint |
||
Issue descriptionIt's expected that resharding perf benchmarks will create regressions. However, what's unexpected is those regressions are bisectable by pinpoint. That's because: 1) Pinpoint doesn't use the sharded benchmarks mechanism as perf waterfall 2) The commandline pinpoint passes to run tests don't include sharding at all.
,
Sep 19
,
Sep 19
Issue 885049 has been merged into this issue.
,
Sep 20
Note those bugs that blamed your CL include the note: "Assigning to nednguyen@google.com because this is the only CL in range" The dashboard saw the range contained a single CL, didn't even bother to run a pinpoint job, and assigned it to you. :) The relevant bit of code seems to be here: https://cs.chromium.org/chromium/src/third_party/catapult/dashboard/dashboard/file_bug.py?rcl=582a06eb78355d29316b9d4a2d3181e725194e52&l=167 Closing as this seems WAI?
,
Sep 20
Ah, very nice. So it's indeed the process of updating the shard during weekend is working well :-) |
||
►
Sign in to add a comment |
||
Comment 1 by nednguyen@chromium.org
, Sep 19