metrics: promote etcd_debugging put_total and delete_total

[hexfusion]

Certain applications such as the Kubernetes apiserver uses txn[1] for Puts which results in a hole for grpc_* metrics calculations. etcd_debugging_mvcc_put_total accurately reflects these metrics for the cluster. Because the etcd_debugging_*[2] namespace metrics have been marked as experimental adoption of these metrics has been stunted. I propose promoting the following. We will maintain backward compatibility with old namespace in 3.4 and plan to depreciate in 3.5

etcd_debugging_mvcc_put_total -> etcd_mvcc_put_total

etcd_debugging_mvcc_delete_total -> etcd_mvcc_delete_total

/cc @gyuho @jpbetz @xiang90 @jingyih

[1] https://github.com/kubernetes/kubernetes/blob/release-1.14/staging/src/k8s.io/apiserver/pkg/storage/etcd3/store.go#L159
[2] https://github.com/etcd-io/etcd/blob/master/Documentation/metrics.md#etcd_debugging-namespace-metrics

[gyuho]

Sounds good. Can you create a PR with CHANGELOG?

[jingyih]

Do we also want to promote etcd_debugging_mvcc_txn_total and etcd_debugging_mvcc_get_total?

[jingyih]

Actually I need some clarification on this. If, for example, client issue a TXN which includes a GET and DEL such as in [1]. After it is applied, all of the following 3 metrics will increase by 1, correct? Just want to make sure these metrics do not confuse users.

etcd_debugging_mvcc_txn_total, etcd_debugging_mvcc_get_total, etcd_debugging_mvcc_del_total

[1] https://github.com/kubernetes/kubernetes/blob/df8cd112a7369f7904682074588a4687cabc2846/staging/src/k8s.io/apiserver/pkg/storage/etcd3/store.go#L194