Better Statuses Kubernetes Executor

**Description**

When running Airflow with the Kubernetes Executor, it will be great to provide better status regarding the execution of the tasks. The best approach would be to display the latest event for the running Pod.

**Use case / motivation**

When users use the Kubernetes Executor they might choose to attach a volume, secrets, configmaps, change the resources requests, etc... as they have the freedom to do so. However if these objects are not present, or the cluster is currently unable to scale to accommodate the Pod, this task will stay in _queued_ for an indefinite amount of time. Without access to the cluster events the user does not have knowledge what the reason is that this task is not executing. Furthermore the tasks will stay in _queued_ and block the execution of child tasks.

This scenario prompts two action items:

1. The user should be provided with the events for the Pod executing the queue task.
2. The user should also be allowed to set a time limit to kill these tasks, not related to the execution time but rather the queue time. In fact setting _execution_timeout_ in the operator does not kill the task if it's not in _executing_ state.

**Related Issues**

N/A