remove some useless metrics

[XuHuaiyu]

Enhancement

In the TiDB server, there are metrics that observe the requests sent from TiDB to TiKV. Under scenarios where there are many cluster nodes and high request load, the data volume for these metrics can become very large. We can remove some unused metrics or optimize them to reduce the data volume

• Remove tidb_tikvclient_cop_duration_seconds_bucket/_sum/_count.
  Update the corresponding metrics source for the monitoring panel TiDB -> Distsql -> Coprocessor Seconds 999 to tidb_tikvclient_request_seconds_bucket. tidb/59845 client-go/1602
• Remove tidb_tikvclient_request_time_counter.
  The tidb_tikvclient_request_time_counter is replaced by tidb_tikvclient_source_request_seconds_sum in client-go/1556, we could abort the collection of tidb_tikvclient_source_request_seconds_count in the Prometheus configuration.
• Remove tidb_tikvclient_request_counter. client-go/1556
• Abort the collection of tidb_tikvclient_batch_requests_sum/_count in the Prometheus configuration. tiup/pull/2523
• Abort the collection of tidb_tikvclient_batch_pending_requests_sum/_count in the Prometheus configuration. tiup/pull/2523
• Abort the collection of tidb_tikvclient_source_request_seconds_count in Prometheus configuration. tiup/pull/2523