Generally, on the server-side, we use the internal as the VLAN. And for health-monitoring (non-floating self-ip) is used to ensure each device performs its own monitoring independently, even in HA pairs. That way each unit can determine, if the pool member is up / down, useful in failover scenarios.

Egress traffic (outbound)—floating self-ip is used to ensure consistency (only the active units uses the shared floating IP to route traffic).

Here, Option C - tcpdump -i internal port 80 and host 172.16.31 where the ip 172.16.31 is only the internal vlan non-floating self-ip availalble in the option.