nova-cloud-controller charm leaving stale neutron-server entries at haproxy.cfg

FYI I'm deploying kilo using current charm trunks,
at n-c-c units:
#1. /etc/init/neutron-server.override is properly set to manual
(still, remains of /var/log/neutron/neutron-server.log are there
showing errors as per lp#1468918

#2. at haproxy.cfg, neutron-server entries still exist:

frontend tcp-in_neutron-server
    bind *:9696
    acl net_172.20.172.222 dst 172.20.172.222/255.255.248.0
    use_backend neutron-server_172.20.172.222 if net_172.20.172.222
    default_backend neutron-server_172.20.172.222

backend neutron-server_172.20.172.222
    balance leastconn
    server nova-cloud-controller-2 172.20.172.191:9686 check
    server nova-cloud-controller-1 172.20.172.203:9686 check
    server nova-cloud-controller-0 172.20.172.222:9686 check

#3. above #1+#2 obviously put check_haproxy_servers into failure
$ cat /etc/nagios/nrpe.d/check_haproxy_servers.cfg
# check haproxy_servers
command[check_haproxy_servers]=/usr/local/lib/nagios/plugins/check_haproxy.sh