Drop deprecated metric to fix failing test.

Dropped via https://github.com/kubernetes/kubernetes/pull/103793

This change drops pod-centric metrics without a non-empty 'container' label.

Previously we dropped pod-centric metrics without a (pod, namespace) label set
however these can be critical for debugging.

Signed-off-by: Damien Grisonnet <dgrisonn@redhat.com>

jsonnet: Drop cAdvisor metrics with no (pod, namespace) labels while preserving ability to monitor system services resource usage

The following provides a description and cardinality estimation based on the tests in a local cluster:

container_blkio_device_usage_total - useful for containers, but not for system services (nodes*disks*services*operations*2)
container_fs_.*                    - add filesystem read/write data (nodes*disks*services*4)
container_file_descriptors         - file descriptors limits and global numbers are exposed via (nodes*services)
container_threads_max              - max number of threads in cgroup. Usually for system services it is not limited (nodes*services)
container_threads                  - used threads in cgroup. Usually not important for system services (nodes*services)
container_sockets                  - used sockets in cgroup. Usually not important for system services (nodes*services)
container_start_time_seconds       - container start. Possibly not needed for system services (nodes*services)
container_last_seen                - Not needed as system services are always running (nodes*services)
container_spec_.*                  - Everything related to cgroup specification and thus static data (nodes*services*5)

The Kubelet's Service/Endpoints object maintained by the Prometheus
Operator does not have the recommended app label (yet). Therefore we
need to use the old label until a Prometheus Operator version has been
released and integrated in kube-promteheus that does use it.

* Add metrics_path label to kubelet servicemonitor endpoints

* Set kubelet metric_path label in jsonnet

* Add generated kubelet servicemonitor

Starting with Kubernetes 1.6 `kubectl apply` works for TPRs so we can
start using it, therefore using multiple files is no issue.