More context on error on WaitUntilPodAvailable

[lahabana]

When writing tests it's useful to figure why a pod failed to start.

For example we use heavily terratest in our e2e suite on kumahq/kuma.
Some tests are flaky for many possible reasons. When investigating them we have to switch to using:
WaitUntilPodAvailableE and use the error to get the pod again to extract the status. Something like:

err := WaitUntilPodAvailableE(...)
if err, ok := err.(k8s.PodNotAvailable); ok {
   p := GetPod(...)
   Logf("bad status", p.status.Message, p.status.Status)
}
require.NoError(err)

Not only there's a race here (possibly the pod gets healthy right after failing but it's also not super easy.

2 options:

1. add pod.Status.{Reason,Message} to PodNotAvailable error
2. make Pod in PodNotAvailable accessible outside the package.

I have a preference for 1.