them

@penguin86 I would believe to recognize the front three as capacitors.

@MrGamy@toot.kif.rocks @kidney@mastodon.online You need one device that is "publicly" reachable (in the above this is achieved by portforwardings through the NATs) for other wireguard clients to connect to. Be aware that depending on your ISP (internet Service Provider) you may not be able to do that (due to shared public IP, search for CGNAT (carrier grade NAT) for more info). Also consider that you may have only ipv4 or only ipv6 connection on either ends.

@MrGamy@toot.kif.rocks @kidney@mastodon.online If you have no hard requirement for "fully selfhosted at no extra cost", you could also consider using a small VPS with a European hoster. The smallest ones for approx 2€/month are sufficient, and you avoid a lot of (possible) problems by avoiding both the port forwarding through double NAT and the dyndns setup. (Plus, the VPS can double as proxy exit node ;) )

@MrGamy@toot.kif.rocks @kidney@mastodon.online with respect to understanding wireguard itself: in each wireguard config there are two very relevant pieces of configuration: the endpoint ("the other end of the tunnel") and the allowedIPs ("what traffic should go through the tunnel to the other side") The rest is not really wireguard specific, but more general IP networking. For wg clients to talk to each other a simple wg config suffices. If you want more (e.g. using other end as proxy) you also need IP forwarding/routing.

@MrGamy@toot.kif.rocks @kidney@mastodon.online Attempting a high level suggestion based on the information given: one of the wireguard endpoints will need to act as "server", presumably the double Nat + dyndns situation. Configure portforwardings for one UDP through both NATs to whatever is the final endpoint running wireguard behind the double NAT. (You may test this with netcat while wireguard is not running). The other end should then simply be able to use the dyndns and UDP port as target to connect to.