Show error when using invalid pod spec

[gregoryfranklin]

When trying to create an agent with an invalid pod.yaml, GoCD only shows 'waiting for agent'. Checking the logs on the server, it shows

2019-01-11 14:27:01,002 ERROR [155@MessageListener for CreateAgentListener] p.c.g.c.e.k.c.g.c.e.KubernetesPlugin:128 [plugin-cd.go.contrib.elasticagent.kubernetes] - Failed to handle request cd.go.elastic-agent.create-agent
io.fabric8.kubernetes.client.KubernetesClientException: Operation: [create]  for kind: [Pod]  with name: [null]  in namespace: [gocd]  failed.
	at io.fabric8.kubernetes.client.KubernetesClientException.launderThrowable(KubernetesClientException.java:62)
	at io.fabric8.kubernetes.client.KubernetesClientException.launderThrowable(KubernetesClientException.java:71)
	at io.fabric8.kubernetes.client.dsl.base.BaseOperation.create(BaseOperation.java:363)
	at cd.go.contrib.elasticagent.KubernetesInstanceFactory.createKubernetesPod(KubernetesInstanceFactory.java:133)
Caused by: java.net.SocketTimeoutException: timeout
2019-01-11 14:27:01,002 ERROR [155@MessageListener for CreateAgentListener] JMSMessageListenerAdapter:77 - Exception thrown in message handling by listener com.thoughtworks.go.server.messaging.elasticagents.CreateAgentListener@2ad4f87c
java.lang.RuntimeException: Interaction with plugin with id 'cd.go.contrib.elasticagent.kubernetes' implementing 'elastic-agent' extension failed while requesting for 'cd.go.elastic-agent.create-agent'. Reason: [The plugin sent a response that could not be understood by Go. Plugin returned with code '500' and the following response: 'Failed to handle request cd.go.elastic-agent.create-agent']
	at com.thoughtworks.go.plugin.access.PluginRequestHelper.submitRequest(PluginRequestHelper.java:57)
	at com.thoughtworks.go.plugin.access.elastic.v3.ElasticAgentExtensionV3.createAgent(ElasticAgentExtensionV3.java:101)

Without any indication as to what the problem is.

For example:

apiVersion: v1
kind: Pod
metadata:
  name: xxxx
spec:
  containers:
    - name: agent
      image: xxxx
      resources:
        limits:
          cpu: 1
          memory: 4096Mi
        requests:
          cpu: 1
          memory: 8096Mi

Should show the following error message which it would get from kubernetes, but this is not visible anywhere in GoCD.

The Pod "xxxx is invalid: spec.containers[0].resources.requests: Invalid value: "8096Mi": must be less than or equal to memory limit

A possible reason for this is that it seems to take kubernetes 30 secs to return the error (when using kubectl create -f pod.yaml) which is longer than the 10 second timeout in GoCD.

Kubernetes: v1.10.9-gke.0
GoCD: 18.6.0
Kubernetes Elastic Agent Plugin: 1.0.1

[arvindsv]

I think this should be done across elastic agent plugins, and not just for this. But, it is a valid problem.

/cc @prateekbaheti @SanjanaB @sheroy Thoughts?