Conversion function to turn PG result into float64 utilization

Hello there,

I might be wrong in here as I am no expert.
I was trying to add replay_lag metric, so I have added a custom ColumnUsage and a conversion function to output a float64.

case CUSTOMCOLUMNUSAGE:
	thisMap[columnName] = MetricMap{
		vtype: prometheus.GaugeValue,
		desc:  ...,
		conversion: func(in interface{}) (float64, bool) {
			// my conversion here
		},
      }

However I was receiving this error Unexpected error parsing column: ... and I have noticed that dbToFloat64 is used no matter the type of column :

https://github.com/percona/postgres_exporter/blob/7b2ff2c41dbca454209e4d40697a7fcd9e3c8c0a/cmd/postgres_exporter/postgres_exporter.go#L1369

Shouldn't be instead:

value, ok := metricMapping.conversion(columnData[idx])

After I changed it my metric was converted correctly and I could query it on Grafana. Is this correct or am I missing something?

If needed I can make a pull request.

On a side note, if you are interested I can open another issue for adding replication lag to the metrics.

Thanks