While working on adding a new e2e to exercise runtimeConfig options, I realized that our E2Es are slow partially because we have extended intervals for wait events - at 30s. This is unfortunate as the first check isn't even called for that 30s period, and meant that the no-op "wait for object to exist" also took 30s.
This speeds up e2es by removing extra waits, and reducing intervals to allow tests to make progress faster when not blocked on a slow network/very slow cpus, without making things worse in those cases.
We also move creating the executor into the test setup, as it's required for all tests, and we don't currently do per-test namespaces.
While working on adding a new e2e to exercise runtimeConfig options, I realized that our E2Es are slow partially because we have extended intervals for
waitevents - at 30s. This is unfortunate as the first check isn't even called for that 30s period, and meant that the no-op "wait for object to exist" also took 30s.This speeds up e2es by removing extra
waits, and reducing intervals to allow tests to make progress faster when not blocked on a slow network/very slow cpus, without making things worse in those cases.We also move creating the executor into the test setup, as it's required for all tests, and we don't currently do per-test namespaces.