Add auto-batching client

[rbino]

The existing stateful client (or possibly a separate one) can support automatically batching incoming requests, to allow collecting requests from multiple services maximizing TigerBeetle capabilities (see here).

The behaviour can be a client that waits until a batch size is reached or a timeout occurs, whichever comes first.

[lessless]

Hey @rbino,

Thank you for your work on the TB client. I'm just catching up with all its capabilities and and noticed that they moved away document you linked in the description.

Thanks!

[rbino]

Hey @lessless, thanks for the heads up, I'll see if I can find a new reference.

I've also done some experimenting with auto-batching for the TigerBeetle Hackaton at Interledger Summit 2023, you can find a video about this here and the code is here

[rbino]

The native TigerBeetle client already does this automatically now (see tigerbeetle/tigerbeetle#1257), so there's no need to support it at the Elixir client level.

Closing this.

[lessless]

Hey @lessless, thanks for the heads up, I'll see if I can find a new reference.

I've also done some experimenting with auto-batching for the TigerBeetle Hackaton at Interledger Summit 2023, you can find a video about this here and the code is here

Thanks for sharing @rbino ! That's was a great presentation. It nicely matches the TB engineering environment.

I wish I would have a chance to work with it in prod one day :)