Although may be a bit long, but totally worth your time to read through it. https://medium.com/@tracymacding/deep-dive-to-starrocks-io-model-part-1-concepts-ec67044881b2

Great article, how to define the correct number of buckets? Also how to set the correct number pipeline_dop to enable infra tablet parallelism? Thanks

I absolutely love your articles, man! 👍 Keep going, pls 🙂

@Giannis Polyzos speaking of the devil

@Julien David Generally speaking, having each tablet hold about 1GB of compressed data is appropriate. Try not to set pipeline_dop manually — the system will handle it adaptively.