So the data is stored in the internally developed Go database? How reliable is that?

Also, the dataset of that size shall be distributed. Why Cassandra did not fit?

Genuine questions - I am evaluating which tool to pick for my own stuff.