Prow integration test is flaky

[smg247]

pull-prow-integration fails frequently with the following flake:

=== FAIL: test/integration/test TestDeletePod/running-pod (60.04s)
    sinker_test.go:246: Creating client for cluster: kind-kind-prow-integration
    sinker_test.go:263: Creating prowjob: running-pod-f7470c65da881743dda945debe3495a7
    sinker_test.go:267: Finished creating prowjob: running-pod-f7470c65da881743dda945debe3495a7
    sinker_test.go:271: Creating pod: running-pod-f7470c65da881743dda945debe3495a7
    sinker_test.go:275: Finished creating pod: running-pod-f7470c65da881743dda945debe3495a7
    sinker_test.go:286: Wait for sinker deleting pod or timeout in 1 minute: running-pod-f7470c65da881743dda945debe3495a7
    sinker_test.go:306: client rate limiter Wait returned an error: rate: Wait(n=1) would exceed context deadline
=== FAIL: test/integration/test TestDeletePod (0.00s)

Example failing run: https://prow.k8s.io/view/gs/kubernetes-ci-logs/pr-logs/pull/kubernetes-sigs_prow/308/pull-prow-integration/1853470929506537472

[smg247]

/kind bug

[petr-muller]

Interesting, this looks new 🤔

[smg247]

The original description was a real failure, but it is flaky. The actual flake is now in the description.