See the bots on the main waterfall that run browser_tests with plznavigate

https://build.chromium.org/p/chromium.linux/builders/Linux%20Tests?numbuilds=200
https://build.chromium.org/p/chromium.win/builders/Win10%20Tests%20x64?numbuilds=200
https://build.chromium.org/p/chromium.mac/builders/Mac10.12%20Tests?numbuilds=200

In the last 200 runs,
Linux had 13 failures in plznavigate browser_tests. 8 in browser_tests without plznavigate.
Win had 8 failures in plznavigate browser_tests. 2 in browser_tests without plznavigate.
Mac had 9 failures in plznavigate browser_tests. 7 in browser_tests without plznavigate. note i ignored infra failures.


I don't really see this on 
https://build.chromium.org/p/chromium.fyi/builders/Browser%20Side%20Navigation%20Linux?numbuilds=200


The failures seem to be all over the place. Since it's only browser_tests and not content_browsertests, could there be a recently introduced memory bug that is only present in src/chrome & plznavigate code paths?

Camille: assigning to you to investigate/delegate.