The CockroachDB documentation v2.1 has a cryptic description of CPUs with NUMA (Non-uniform Memory Access):

To add more processing power (up to 16 vCPUs), adding more vCPUs is
better than adding more RAM. Otherwise, add more nodes rather than
higher vCPUs will have
NUMA(non-uniform
memory access) implications. Our internal testing results indicate
this is the sweet spot for OLTP workloads. It is a best practice to
use uniform nodes so SQL performance is consistent.

But all mention of NUMA has been removed from the later version of this document recommended-production-settings.md.

There seems to be no recommendations for CRDB regarding NUMA and how to configure Linux. I’ve heard CRDB is not NUMA-aware, but this is not documented nor is the performance impact of running on server with NUMA.