Hi @GradyMurphy7996,
"Do I need to configure the vlan across the entire vxlan environment or just on the the single switch?"
A VXLAN topology is usually built based on the CLOS topology, this is Spine and Leaf Switches.
- Spine switches route the traffic between Leaf switches. They only do Layer 3 Routing (most commonly OSPF or ISIS) and are usually the BGP EVPN Route Reflectors of the Leaf switches.
- Leaf switches are the ones you connect your Servers to. You configure the VLAN and VXLAN related parameters (like L2VNI and L3VNI) here.
These are the ones who receive the original Ethernet frame from a Source Server "encapsulate it in VXLAN" and forward them to the Spines who just forward them to the correct destination Leaf for this traffic to be now "decapsulated" and forwarded to the destination Server.
As a conclusion, you only need the new vlan & vxlan related commands on the respective Leaf Switches where you have servers on the same "VXLAN" domain.
"Also, they will need L3 connectivity to reach out to other resources, so how do I establish that as well?"
Usually, there is one or two Leaf Switches (when there are two it is better to have them on a vPC cluster) acting as "Border Leaf" Switches.
This "Border Leaf" can be the same Switch acting as "Leaf" where you have your Server connected to.
On Border Leaves is where communication to the "outside the fabric" resides.
Most likely the Border Leaf receive the routes via a Dynamic Routing Protocol and redistribute those into BGP EVPN for the fabric to have routing to "outside the fabric"
I personally think there are a lot of similarities between MPLS L3VPN and VXLAN BGP EVPN... So if you have a good grasp of the former it will be easier for you to understand the latter.
In any case, this document is a good starting point:
VXLAN Network with MP-BGP EVPN Control Plane Design Guide
Cheers.
