All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=808656

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=2fdc7150d21fe34f2ff2d62b8823559a637c5b6a4fb097a459380062a1228bb6


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

android-nexus5X
android-nexus6
android-one

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/16c7c62e840000

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14af01fc440000

Re-kicking the bisect. Looks like it reproduces, but didn't get to the culprit.

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/14af01fc440000

Enable SpeculativePreconnect Field Trial Testing by alexilin@chromium.org
https://chromium.googlesource.com/chromium/src/+/8442e044284223377ccd75e4f1b519c5656d4d8d

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

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/12c585b9440000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/12c585b9440000

predictors: Skip dns preresolution if a proxy is configured by alexilin@chromium.org
https://chromium.googlesource.com/chromium/src/+/b6e03254a007461585b9b484e620893630f127a8

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

The metric has recovered after https://crrev.com/c/958913

Thanks for the improvement from #7!

This bug has 11 graphs, and it's worth pointing out that there is a regression after this one in most graphs. However most timeToFirstMeaningfulPaint_avg graphs are not recovered at all. So I don't think this should be considered fully fixed. That is, I think we would see an improvement in these graphs if the field trial experiment was reverted.

Actually it seems those graphs haven't caught up to that CL, so let's wait until they do to see if they recovered.

Sorry for being too eager to close this bug, you're right I had to wait until all graphs caught up to that CL. I think all of them have finally recovered from this regression.