This issue has been available for more than 365 days, and should be re-evaluated. Please re-triage this issue.
The Hotlist-Recharge-Cold label is applied for tracking purposes, and should not be removed after re-triaging the issue.

For more details visit https://www.chromium.org/issue-tracking/autotriage - Your friendly Sheriffbot

Ken, can we add a new failure mode that if harness crashes during running a test, then that test is marked as fail?

This would be difficult. If the GPU process crashes in a way that the test doesn't notice it, then at some point a lost context should be delivered, but possibly on the next test. While it would theoretically be possible to add code that asserts that the GPU process crash count is the same at the beginning and end of the test, this could be flaky and racy.

At this point this particular test is passing well, and subsequent tests also pass, indicating to me that the GPU process isn't crashing any more. We have seen other tests fail reliably if the GPU process crashes, so I'm hesitant to add more code to the test harness. If this becomes a problem again in the future let's open a new bug.