Hi there. We just went through this very same issue. After quite a bit of trial, error and tcpdump, we were able to narrow it down to the "interface ignore wildcard" option in /etc/ntp.conf. Removing this line and restarting the ntp service fixed the issue for us. Note that removing this line causes ntpd to bind to 0.0.0.0:123/udp (but not answer NTP queries using the default config) which may or may not be acceptable in your environment. Our network ACLs prohibit external connections to this port. Prior to effecting this change running "ntpq -p" would show all peers stuck in the "INIT" state forever.
We're now trying to figure out how to:
a) report this as a bug in the virtual appliance AMI
b) run a command on the EC2 instance on first boot to remove that line from the ntp.conf and restart the service as a workaround in the meantime.
Any tips?
Search for "prevents opening matching addresses" in the following manpage for more info:
http://manpages.ubuntu.com/manpages/trusty/man5/ntp.conf.5.html
HTH someone else.