All graphs for this bug:
  https://chromeperf.appspot.com/group_report?bug_id=821399

(For debugging:) Original alerts at time of bug-filing:
  https://chromeperf.appspot.com/group_report?sid=5dcc4d30fdc3ee74e98e209a39c838de2a95733794832fd3249b32671cd50c96


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

chromium-rel-win7-dual

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/14934c96440000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/14934c96440000

Introduce InputDeviceInfo interface by c.padhi@samsung.com
https://chromium.googlesource.com/chromium/src/+/636394fb5626827fd9c15e0a60df4ce9fe8af6f0

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

📍 Pinpoint job started.
https://pinpoint-dot-chromeperf.appspot.com/job/17bf058e440000

📍 Found a significant difference after 1 commit.
https://pinpoint-dot-chromeperf.appspot.com/job/17bf058e440000

Introduce InputDeviceInfo interface by commit-bot@chromium.org
https://chromium-review.googlesource.com/c/chromium/src/+/942961/7

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

guidou@: The above is the result of the test I ran to test https://chromium-review.googlesource.com/c/chromium/src/+/942961 with the current configuration, after following the steps in http://g.co/ChromePerformanceRegressions. It doesn't seem to cause any regression. Is my understanding correct? I am not really familiar with these tests.

Do you think the CL in question might cause any regression as it only adds a new interface?

I also found a CL https://chromium-review.googlesource.com/c/chromium/src/+/959945 that seems to fix query-selector-all tests regression caused by an earlier CL. This CL was merged today.

Bit early to say for sure but looks like the graph recovered after the CL mentioned #6. Lets leave another day then mark Wontfix if we're still good.

I don't think https://chromium.googlesource.com/chromium/src/+/636394fb5626827fd9c15e0a60df4ce9fe8af6f0 can cause this performance regression, but let's follow hjd@'s advice.

Graph looks good :) Closing.