The article fixates excessively in throughput, as if was the only factor that mattered for choosing a data store. It isn't, since Kafka isn't necessarily about

The article fixates excessively in throughput, as if was the only factor that mattered for choosing a data store. It isn't, since Kafka isn't necessarily about scaling, but also focuses on immutability, architectural flexibility etc, all of which a RDBMS might not offer.