CDK grafana dashboard using missing/renamed metrics

[Sérgio Manso]

The grafana dashboard available via relation[1] is making use of metrics that are either renamed or removed from upstream.
The following metrics were identified as problematic:

- apiserver_request_latencies_sum (got renamed apiserver_request_duration_sum)
- apiserver_request_latencies_count (got renamed apiserver_request_duration_count)
- apiserver_request_count (got renamed apiserver_request_total)
- apiserver_request_latencies_bucket (got renamed apiserver_request_duration_bucket)
- admission_quota_controller_adds (renamed to workqueue_adds_total) See PR [2]
- admission_quota_controller_queue_latency_sum (renamed to workqueue_work_duration_seconds_sum)
- admission_quota_controller_queue_latency_count (renamed to workqueue_work_duration_seconds_count)
- etcd_request_* and etcd_helper_* removed from upstream. See PR [3]

As a workaround, these metrics were manualy fixed in Grafana.

[1] https://github.com/charmed-kubernetes/charm-kubernetes-master/blob/master/templates/grafana/autoload/kubernetes.json
[2] https://github.com/kubernetes/kubernetes/pull/71300/files
[3] https://github.com/kubernetes/kubernetes/pull/74418/files#diff-5a0d6b0e34ff7529a142eec3648065b523e9c69fba321eefd7eed5636b27ecd9L45-R45

[Sérgio Manso (sergiomanso)]

adding a fixed grafana dashboard as an example

[Chris Johnston (cjohnston)]

PR: https://github.com/charmed-kubernetes/charm-kubernetes-control-plane/pull/287