3 ms·

Yeah, you can process 99% of tabular workloads with that. I generally advise my clients to work on a single node before attempting to scale out.