Right now coredumps in unit tests are only collected during a crash, and not when the unit test times out. Setting up the timeout core dump collection logic is going to require collaboration with EngFlow and likely not be completed until at earliest next week.

To workaround this issue, I would recommend reproducing timeout issues locally with a bash loop of bazel run

that you leave running on your workstation in the background for a while. If this isn't viable please send me a DM on slack and we can figure something else out.