Skaffold reports pod is ready based on phase not readiness state

[briandealwis]

In CC-VSC, I saw in the output:

Status check started
Resource pod/mongodb-deployment-587b66548b-nmgsh status updated to In Progress
Resource pod/nodetodo-deployment-5dcbd4795b-94sqb status updated to In Progress
Resource pod/mongodb-deployment-587b66548b-nmgsh status completed successfully
Resource deployment/mongodb-deployment status completed successfully
Resource deployment/nodetodo-deployment status updated to In Progress
Resource pod/nodetodo-deployment-5dcbd4795b-94sqb status completed successfully

So pod/nodetodo-deployment-5dcbd4795b-94sqb appeared to be running. But kubectl disagrees:

$ kubectl get all
NAME                                       READY   STATUS    RESTARTS   AGE
pod/mongodb-deployment-587b66548b-nmgsh    1/1     Running   0          92s
pod/nodetodo-deployment-5dcbd4795b-94sqb   0/1     Running   0          92s

NAME                        TYPE           CLUSTER-IP       EXTERNAL-IP   PORT(S)        AGE
service/kubernetes          ClusterIP      10.96.0.1        <none>        443/TCP        16d
service/mongodb-service     ClusterIP      10.98.37.129     <none>        27017/TCP      92s
service/nodetodo-external   LoadBalancer   10.104.156.208   <pending>     80:32354/TCP   92s

NAME                                  READY   UP-TO-DATE   AVAILABLE   AGE
deployment.apps/mongodb-deployment    1/1     1            1           92s
deployment.apps/nodetodo-deployment   0/1     1            0           92s

NAME                                             DESIRED   CURRENT   READY   AGE
replicaset.apps/mongodb-deployment-587b66548b    1         1         1       92s
replicaset.apps/nodetodo-deployment-5dcbd4795b   1         1         0       92s

$ kubectl describe pod/nodetodo-deployment-5dcbd4795b-94sqb
...
  Warning  Unhealthy  32s (x4 over 4m38s)  kubelet            Readiness probe failed: Get "http://172.17.0.5:3000/todo": EOF
  Warning  Unhealthy  32s (x3 over 4m32s)  kubelet            Liveness probe failed: Get "http://172.17.0.5:3000/todo": EOF
  Normal   Killing    32s                  kubelet            Container nodetodo failed liveness probe, will be restarted
  Warning  Unhealthy  32s                  kubelet            Readiness probe failed: Get "http://172.17.0.5:3000/todo": dial tcp 172.17.0.5:3000: connect: connection refused

Eventually the output view added:

Resource pod/nodetodo-deployment-5dcbd4795b-94sqb status failed with Readiness probe failed: Get "http://172.17.0.5:3000/todo": EOF

I suspect our status checks are reporting the pod phase and ignoring the ready state — the same problem as in #5308.

(This caused some head-scratching trying to understand why the app wasn't running properly.)