We use SQL data sources and then Meta Sources to scale the raw data into meaningful data. To avoid Mango creating data points while communication was lost from the data logger we have Mango set to read on a timestamp change only.
On the SQL data source we have a data point just for the timestamp, the other data points use a time override column name and on the meta source we use the timestamp data point for the context update.
Of course this will only work if your logging device creates its own timestamp or your SQL database records the date and time of the data insertion into the table.
I hope this helps