Greetings: I am configuring a cluster using RKE2. I understand that nodes are for the control plane and that agents are for running application workloads. Is there any best-practice advice on mixing/matching on the same node? also what would the memory requirements be? I have only 3 nodes which have blocks of memory that i can adjust. Currently they have 4gb and this is for a dev cluster. Interested in the optimal setup for this small size? 1 node/2 agents, 2 nodes - 1 agent? thanks for any insight!

Server nodes run control-plane components, etcd, and kubelet. Agent nodes run only the kubelet. All nodes may run workload pods; it's up to you to decide if you want to prevent that by tainting the servers.

thank you! understood

1 server and 2 agents is probably the best use of resources for a dev cluster that does not need HA

yes! that is what i was thinking also. Appreciate your help!