Worker node sizing

Does anyone here have any gather metrics for sizing worker nodes? I know the recommendation within AWS is an M5l/n node, but how many sessions are people able to sustain over one? We’re setting up some self-managed workers within AWS that will predominantly be used for RDP sessions and were trying to gauge how many nodes we may need to support traffic. Our current access pattern runs around 600+ sessions per day distributed across different VPCs which will each have separate worker nodes. Any advice would be appreciated.