In our project we had to build a VPN to get through to computers residing behind NAT. I never did it before. While looking for suitable software I came across WireGuard which claimed to be very simple.
After some reading I indeed was able to setup a server with 8 lines long config file and a client (that was behind NAT) with 9 lines long config.
The link worked perfectly in both directions. After that for obvious reasons I don't want to touch any alternatives. WG is on its way to the mainstream Linux kernel, but it is not there yet.
The protocol was formally verified and technical whitepaper exists as well. However the website claims WireGuard shouldn't be "relied on".
How risky is it, from information security point of view, to use WG in production in its current state? What are potential troubles?
Update: As of 2020 WireGuard is inside mainstream Linux kernel, so quite production ready.