I am pulling data using oracle-tap from an external database which has some rare UTF-8 encoding errors. Once every few million rows, some weird character is found and meltano crashes with the following message:

UnicodeDecodeError: 'utf-8' codec can't decode byte 0xe2 in position 1999: unexpected end of data

I wonder how to approach this situation. • I think one way would be to force meltano tap to transport raw data exactly as is, without any checking. But I wonder whether this is possible since data travels in json. • Another way would be to perhaps catch an exception and report the faulty record in an error log, however I see no reason why this should break the ETL. The faulty row could be either omitted, or written empty, or written with faulty characters replaced. However, for the moment I don't know how to achieve any of the above. The process just breaks. Any ideas? Does Meltano have universal mechanisms for this, or best practices to deal with faulty data? Or is everything depending on the plugin?