I have a Lightsail instance where IPv6 is activated. I want to communicate using IPv6 with other EC2 resources in my default VPC outside Lightsail (in the same availability zone). I have activated VPC Peering in the Lightsail console in my Lightsail account. In my default VPC I can see that an entry has been added to the route table that forwards traffic destined to IPv4 addresses belonging in the Lightsail VPC to the pcx-* peering connection. However, a corresponding IPv6 entry is not automatically added. Sending some ICMPv6 ping packets to the Lightsail instance from my EC2 instance, at the receiving Lightsail instance I can see in Wireshark that the TTL in the header is one less than it is at the sender side, indicating that this is routed through an external router outside my VPCs and not using the VPC peering, so I assume this results in extra data traffic charges. If I manually add a corresponding IPv6 forwarding rule in the route table, the TTL is now not decremented anymore and hence indicates the traffic flows through the VPC peering. Unfortunately, I find no way to do the same way in the other direction (traffic from Lightsail into my default VPC), since the route table in Lightsail's VPC cannot manually be modified as far as I know. When sending ICMPv6 from Lightsail into EC2, I see the TTL has been decreased, hence not flowing through the VPC peering I guess, and therefore costs extra data transfer charges (I guess?). It thus seems the activating VPC peering switch in the Lightsail console forgets to add corresponding IPv6 entries in the (hidden) route table in the Lightsail VPC. Can this be fixed somehow or did I miss something? If not, see this as a feature request that I guess shouldn't be that hard to implement.
