Example: https://ci.chromium.org/p/chromium/builders/luci.chromium.ci/Mac10.12%20Tests/18654

Findit seems not perform well on analyzing very broken builds (too many failed steps, too many failed tests, etc). It'll trigger too many swarming tasks(if too many failed steps), or try job callback tasks become too big (if too many failed tests).

We need to think about something to better handle these.

Easiest solution is to bail out on those builds? Or limit the number of tasks/size of test list to analyze the failure.