Misleading message is logged during retries if the service not exists

[gyuuu]

When the configured service not exists, the logged messages can be misleading.

When service discovery is set up to use a service which not exists, the following message get logged during the retires:
Couldn't connect to the Kubernetes master, [%s] retrying in %s seconds...

Based on the error message, I assumed that the API server is unaccessiable from the pod.

After 10 retries I got the real error message:
Failure executing: GET at: https://kubernetes.default.svc/api/v1/namespaces/test/endpoints/test-service

Configuration:

config.getNetworkConfig().getJoin().getKubernetesConfig().setEnabled(true)
                    .setProperty("namespace", "test")
                    .setProperty("service-name", "test-service");

It's not a big deal at all, however it can lead to a few hour long debug session if your application is killed by the readiness check before the real error message get logged.