ESQL: Per column profile info for ValuesSourceReaderOperator

[ioanatia]

This is the current format for the profile for ValuesSourceReaderOperator:

          {
            "operator": "ValuesSourceReaderOperator[fields = [content, content.keyword, namespace, namespace.keyword, redirect, redirect.keyword, title, title.keyword, _id]]",
            "status": {
              "readers_built": {
                "_id:column_at_a_time:null": 27,
                "_id:row_stride:BlockStoredFieldsReader.Id": 18,
                "content.keyword:column_at_a_time:BytesRefsFromOrds.Singleton": 18,
                "content:column_at_a_time:null": 27,
                "content:row_stride:BlockSourceReader.Bytes": 18,
                "namespace.keyword:column_at_a_time:BytesRefsFromOrds.Singleton": 18,
                "namespace:column_at_a_time:null": 27,
                "namespace:row_stride:BlockSourceReader.Bytes": 18,
                "redirect.keyword:column_at_a_time:BytesRefsFromOrds.Singleton": 18,
                "redirect:column_at_a_time:null": 27,
                "redirect:row_stride:BlockSourceReader.Bytes": 18,
                "stored_fields[requires_source:true, fields:1, sequential: false]": 27,
                "title.keyword:column_at_a_time:BytesRefsFromOrds.Singleton": 18,
                "title:column_at_a_time:null": 27,
                "title:row_stride:BlockSourceReader.Bytes": 18
              },
              "values_loaded": 1730,
              "process_nanos": 215099458,
              "pages_received": 27,
              "pages_emitted": 27,
              "rows_received": 366,
              "rows_emitted": 366
            }
          }

It would be great if we could get process_nanos per each reader, instead of the total.
Then we would know exactly which reader is slow, instead of trying different combinations.

[elasticsearchmachine]

Pinging @elastic/es-analytical-engine (Team:Analytics)