Query processing speed limit is highly dependent on the actual data and the specific query. Ingestion rate is a scalable thing, and BTW ingestion likely becomes slower as the more sophisticated compression and the smarter indexes are used.

I don’t understand why we cannot say that the limit is the same without saying what is it specifically. The point was that there are just no factors (at least I don’t know them) that could differentiate either ClickHouse, Druid and Pinot. I’ve given an example of differentiating factor that makes InfluxDb inherently slower, so InfluxDb is going to have a different “limit”.