Anyway, I replaced the HDMI cables with the latest greatest that supports 8K and shit and it has never happened since. Didn't know HDMI cables could just go bad like that
@pomstan@david@pwm@i@phnt@mint@hfaust@sun What good is having an allocation if next time the ISP router reboots you get a completely different one? This is normal behavior for a bunch of ISPs
@pomstan@david@pwm@i@phnt@mint@hfaust@sun I've been hearing that IPv6 is the future since the first World IPv6 day (which the ISP I worked at participated in!) and still it has barely gained traction in the West except on mobile networks. Fortunately mobile phones are a huge use of the internet today so there is some significant traction there, but not really for wired service (cable/fiber/DSL). And often those who do offer IPv6 won't give you a stable allocation so you can't properly leverage a /56 for your own networks. You end up having to use NPTv6 which is basically NAT (sorta, it just replaces half of the address and replaces it with whatever your ISP is currently giving you)
Still not a lot of movement on IPv6 with some of the largest websites/providers. Amazon is now trying to push customers to IPv6 and charging more for IPv4 addresses, but they aren't deploying IPv6 for all their own services ... :sigh:
@sun@mint@david@pwm@i@phnt@hfaust We should never allow an unauthenticated request (not one of our users) make external API calls that can hang and delay our ability to respond.
Kinda think the smart thing to do would be to deny the first attempt at authenticated fetch because we don't know them and insert an Oban job to fetch their instance actor. Future requests could then begin working, possibly within a second or two. Also will let us prevent duplicates / hammering their server with requests by limiting unique jobs to once per domain like every 5 mins or something