was my understanding wrong?

[jfbaro]

Sorry,

Now I see that the Serverless that Neon is currently exposing to the customer is not the same one that I thought. For me, Serverless Database is something in the sense of pay as you use (per transaction).

Something like:

Each 1.000.000 (8kb) (IO) reads will cost you USD X.XX
Each 1.000.000 (8kb) writes will cost you USD X.xx (Bitemporal cost uses 4x more data per write).

All the complexity about how indexes are used, how IO is used, how high availability is provided, how many instances are running or how much "RAM" is being allocated, should all be abstracted away from the user.

Can any one please educate me on what is the "serverless" offered by Neon?

BTW, I like the idea of having more PostgreSQL options available.

Thanks