Make deploy.RunPod wait for pod readiness rather than running phase

[nadiamoe]

Description

This PR changes the WaitPodRunning function to WaitPodReady, which checks whether the pod has the Ready condition set to True instead of simply being in the Running phase.

If a pod does not contain any container with a readiness probe defined, WaitPodReady should behave exactly the same as WaitPodRunning did. However, if one or more containers have a readiness probe, WaitPodReady will wait until they are satisfied.

This is helpful to reduce flakiness on certain tests which previously attempted to initiate a connection to a pod once it was running, but before the pod had time to actually get started. By attaching a readiness probe to those pods, this scenario can be avoided, as deploy.RunPod will wait or error until the readiness probe passes.

This PR also adds a helper WithHTTPReadinessProbe function to ContainerBuilder, that attaches an HTTP GET readiness probe to a container with reasonable defaults.

Checklist:

• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation
• I have added tests that prove my fix is effective or that my feature works.
• I have run linter locally (make lint) and all checks pass.
• I have run tests locally (make test) and all tests pass.
• I have run relevant integration test locally (make integration-xxx for affected packages)
• I have run relevant e2e test locally (make e2e-xxx for disruptors, or cluster related changes)
• Any dependent changes have been merged and published in downstream modules
  

[pablochacin]

I think the need for the BuildNginxPodWithoutProbes() fixture for the disruptor tests a design smell. The idea of introducing a readiness probe in the test fixtures was to avoid the flakiness of certain tests, but it introduces more subtle problems as the disruptor tests will fail if this non-intuitive fixture is used.

The reason for these potential test failures is that the tests apply the disruption to all requests to a target pod or service. This includes the readiness/liveness probes if any, unless explicitly excluded. But which probes to exclude may not be obvious and depended on the fixture's definition

I think that either we reverse the situation and make the alternative BuildxxxWithProbes fixture or we address the problem of automatically excluding the pod's probes from disruptions.