Hello! Can somebody share some performance characteristics for realtime tables? How fast can server ingest from kafka, how to scale via partitions etc? Thanks

Hello, some stats from production deployments: a 16core 64GB instance has handled 100k events per second (can go higher/lower depending on schema, transform function, etc).

That's great. What is scaling options? More real time servers for each kafka partition? Customer is targeting to peak 1-1.2 mio/s with relatively simple structure, cca 16 columns. Is this capability linearly scalable or whats is pattern here? Thank you

what is mio?

Million

For scaling you can have multiple realtime servers, and each consuming a subset of partitions

👍👍 thank you

@User you can watch this one: