Other than keeping an eye on their changelog or waiting until it breaks, I don’t think you can do anything about that. I do have automatic update, but the config rarely changes from my experience.

I also recommend it.

It works with containers so I can create a setup where requests sent from the container goes through the VPN. I use it for my Redlib setup to bypass rate limiting by rotating its IP regularly. Unless you have your host to route all traffic through a certain node, it should work independently from Tailscale.

Anubis, though I always had it before I removed Cloudflare.

CLOUDFLARE IS NO MORE FOR MY NETWORK

Soon I’ll drop Cloudflare for my public services too

Definitely more than a year! If you have tried it in the past, you probably dropped it either because you used it before the revival, or the UI looked really old. At least that was what I did.

It works pretty well despit having 30k+ music files read over rclone, though I am the only user. It also has a web client, though it looks a bit old. I use Symfonium on Android and Feishin on Desktop since it provides OpenSubsonic API.

One major reason why I have Ampache as a separate server is that they support smart playlist, which wasn’t well supported on Jellyfin. Navidrome also supports smart playlist, but you couldn’t edit on the web.

sops-nix + rootless podman turns out to be much trickier than I imagined. Spent like 2 days over this shit just to get it in the central config when I could have just manually loaded the config files and change the permission… I eventually solved it by running rootlesskit in the activation script to copy the decrypted file into a temporary folder and changing the permission to the correct sub-user. Not worth the time though.

It’s actually not just one single infinitely scrollable workspace; you can do most of the stuff you usually do in any other tiling window manager. Multiple workspace system is still there, you can have multiple windows in a single column.

What makes niri really good is that you can have multiple full-screen window in a single workspace.

I was talking more about whether they can personally tolerate it or not. I thought Factorio over Wi-Fi would be okay even with the inevitable latency, but it was slightly off in a way that I simply could not continue. Meanwhile, I’ve seen people playing ranked games of Rainbow 6 Siege with a similar setup.

If you haven’t used Sunshine to play games yet, I would first try it out with whatever equipments you have before going all-in. It sounds fucking cool on paper, but the whole experience wasn’t all that great for me. Not the Sunshine’s fault, but the games I play are very latency sensitive that it was barely playable.

Personally, if the games play well, I would just go for it.

In terms of setup, Caddy is a lot simpler in syntax, but you will find more tutorials for Traefik and it has better integration with Docker. You can add labels to a container and Traefik uses that as config, whereas in Caddy, you need to set up both the container and the config file. If you want to drop a service, then it is easier in Traefik for this reason. But with decent Nix code, you can basically replicate this in Caddy. Once you set them up, they’re pretty much the same. I’ve seen some people saying Traefik is faster, but realistically, I don’t think it’s meaningful.

The three letter agencies?

Suspiciously nut-shaped training dataset:

Same. There are some tracks and albums I don’t like, but I won’t delete them. Another reason to use smart playlist, I can just put them into “Not My Style” playlist and it’s magically gone from my main list.

Try Ampache! I host 75k files with it.

Item Count: 74939 | Duration: 5274:37:36

Well, I don’t actually play all of them in a straight line; it’s more of an archive. Still, my main playlist is few thousand songs long.