All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=804314

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=609becc160bf79104ffec8a4a9aa63d0da659078083e63d37252075989124ae7


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

android-webview-nexus6

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12f219a8840000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/12f219a8840000

Roll src/third_party/catapult/ b4826a528..2c47ed7f0 (1 commit)
By catapult-deps-roller@chromium.org · Wed Jan 17 13:36:25 2018
chromium @ eedb5c3144bf2c346bf239f79430db4d4d97a096

Understanding performance regressions:
  http://g.co/ChromePerformanceRegressions

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/12f219a8840000

Juan: The only CL in this roll is "[Telemetry] Migrate snap_page to new browser API". Can you take a look?

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12ab6230c40000

Does not seem to be related to my CL. Rather pinpoint jobs were failing and a "change" in the failing pattern happened around the time when pinpoint started running that roll. I've re-kicked the bisect.

+simonhatch to triage, pinpoint should probably not include errors due to isolate/swarming issues when deciding whether a CL really affected the test/metric behavior.

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/12ab6230c40000

Thanks, filed a bug here: https://github.com/catapult-project/catapult/issues/4383

Looks related to the builder migration as well. ( Issue 828778 )

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14c001b8c40000

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/14c001b8c40000

Should I expect this bisect to work now? Any idea why it keeps failing?

+dtu

These are failing with build errors, and these commits are from Jan 17. Guessing this is the builder migration? How long do we keep those cached builds around?

Yes, looks like I've made a mistake with the builder migration. I've deployed a fix and will re-run this job now.

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/15878c24c40000

😿 Pinpoint job stopped with an error.
https://pinpoint-dot-chromeperf.appspot.com/job/15878c24c40000

Metrics have recovered, so I'm WontFix'ing this.

+simonhatch still, let me take this chance to propose some pinpoint feature requests. 

Last job in #17 failed again with "SwarmingExpiredError".

For cases like this I would love to have a "continue bisect" button (i.e. retry those expired swarming tasks) rather than "retry" the whole job from the beginning. Even better if there is no button. And pinpoint just retries when bots are no longer overloaded.

Alternatively (or in addition to) if I'm retrying the job, maybe I want to tweak (usually narrow down) the commit range to save time?

Finally, I just noted that there is no easy way to get from the pinpoint job to the dashboard graph where the alert got triggered (although the info seems to be there in the "tags" job argument). Could that get linkified?

:)

Sounds good! Filed 2 feature requests to continue the discussion:

https://github.com/catapult-project/catapult/issues/4395
https://github.com/catapult-project/catapult/issues/4396