I've seen @davidbaron do the detective work to realize that one of our test machines in CI had bad memory way too many times. (Lots of random failures/crashes in tests all on the same machine...)
2
2
There's a bunch of history in bugzilla.mozilla.org/show_bu… . But now that the test machines are "in the cloud", I don't think we expose a useful concept of machine identity that we can look at to figure this out anymore.
The other question is what portion of crashes are bad memory.
Jan 14, 2019 · 1:52 PM UTC
1
2

