Add monitoring metrics for column shards

[SloNN]

For a common approach to monitoring with datashards, the following metrics should be added:

table.columnshard.used_core_percents
table.columnshard.read.bytes
table.columnshard.read.rows
table.columnshard.write.rows
table.columnshard.erase.rows
table.columnshard.erase.bytes

That metrics must behave the same way as datashard metrics