Flaky E2E test: ScaleOutFrom1To3

[ivanvc]

The CI fails with:

 === RUN   TestScaling/ScaleOutFrom1To3/ScaleOutFrom1To3/verify_member_list
    e2e_test.go:173: Found unpromoted learner after scaling completed: member etcd-scaleoutfrom1to3-2 (17278970040244508223)
--- FAIL: TestScaling (80.35s)
    --- PASS: TestScaling/ScaleInFrom3To1 (40.07s)
        --- PASS: TestScaling/ScaleInFrom3To1/ScaleInFrom3To1 (40.07s)
            --- PASS: TestScaling/ScaleInFrom3To1/ScaleInFrom3To1/scale_to_1 (10.01s)
            --- PASS: TestScaling/ScaleInFrom3To1/ScaleInFrom3To1/verify_member_list (0.03s)
    --- FAIL: TestScaling/ScaleOutFrom1To3 (40.28s)
        --- FAIL: TestScaling/ScaleOutFrom1To3/ScaleOutFrom1To3 (40.28s)
            --- PASS: TestScaling/ScaleOutFrom1To3/ScaleOutFrom1To3/scale_to_3 (30.16s)
            --- FAIL: TestScaling/ScaleOutFrom1To3/ScaleOutFrom1To3/verify_member_list (0.03s) 

Refer to: https://prow.k8s.io/view/gs/kubernetes-ci-logs/pr-logs/pull/etcd-io_etcd-operator/297/pull-etcd-operator-test-e2e/2028548466015211520#

[ahrtr]

@nwnt One of the e2e test added in #298 is pretty flaky. Could you take a look? thx

If you look at the recently PRs, they are constantly failing.