[frr/bgpd] fix frr VXLAN EVPN configuration for unified config

During debugging of VXLAN EVPN using unified vs split configuration, the
same BGP configuration is made with one exception, the setting of
```
no fpm use-next-hop-groups
```
Is forcibly set for all FRR instances.

This change was introduced in PR #12852 when switching to the new
fpm dataplane plugin.

When running `vtysh -c "config" -c "no fpm use-next-hop-groups"`
it can be seen the VTEP immediately comes online and traffic
flows as expected.

This adds the option as is present in dockers/docker-fpm-frr/frr/zebra/zebra.conf.j2
controlled by the DEVICE_METADATA nexthop_group option.

Signed-off-by: Brad House (@bradh352)

src/sonic-frr-mgmt-framework/templates/frr/frr.conf.j2

@@ -10,6 +10,15 @@
 !
 agentx
 !
+{% if ( ('localhost' in DEVICE_METADATA) and ('nexthop_group' in  DEVICE_METADATA['localhost']) and
+        (DEVICE_METADATA['localhost']['nexthop_group'] == 'enabled') ) %}
+! enable next hop group support
+fpm use-next-hop-groups
+{% else %}
+! Uses the old known FPM behavior of including next hop information in the route (e.g. RTM_NEWROUTE) messages
+no fpm use-next-hop-groups
+{% endif %}
+!
 ！Add fpm address for zebra
 fpm address 127.0.0.1
 ！