ClickHouse suits well for string logs, especially if a log message have some fixed structure which can be mapped to separate columns. And most logs have some number of predefined fields which are easy to map to columns (e. g. timestamp, IP, request time, e. t. c.). You can store free-form JSON which doesn't have per-defined schema in a string column, but ELK would be probably better in this case.

And then you can query logs using all power of SQL.