UPSTREAM: <carry>: Still report panic stack traces

A consequence of the previous performance change was losing full
stack traces on runtime panics. Create a new method that internal
locations that must capture panics use.

UPSTREAM: <carry>: Do not capture the full stack trace

Stack traces are expensive and this is not used within our tests
in a meaningful way. This reduces startup time of openshift-tests
from 1.2s to 0.3s. It reduced total e2e run time from 10k core
seconds to 5k core seconds.

UPSTREAM: <carry>: Do not capture the full stack trace

Stack traces are expensive and this is not used within our tests
in a meaningful way. This reduces startup time of openshift-tests
from 1.2s to 0.3s. It reduced total e2e run time from 10k core
seconds to 5k core seconds.