exu

You're right with the origin. codename or n in short form is any-version. ${distro_codename} won't match that, as it contains the codename for your distro release, like bookworm for Debian 12.
With any-version the repo owner's basically saying you can install this regardless of your distro version or they handle it on their end somehow.

Try just using the origin instead, like this.

"origin=cloudsmith/caddy/stable";

Unattended Upgrades only checks and updates programs in repos it knows about. As you found out, you'll need to add the custom repository to the Origins pattern in 50unattended-upgrades.

You can find a list of all repositories and their data using apt policy

Here are the custom repositories I have on one of my servers:

 500 https://repo.zabbix.com/zabbix/7.0/debian bookworm/main all Packages
     release v=12,o=Zabbix,a=zabbix,n=bookworm,l=zabbix,c=main,b=all
     origin repo.zabbix.com
 500 https://repo.zabbix.com/zabbix/7.0/debian bookworm/main amd64 Packages
     release v=12,o=Zabbix,a=zabbix,n=bookworm,l=zabbix,c=main,b=amd64
     origin repo.zabbix.com
 500 https://pkgs.tailscale.com/stable/debian bookworm/main all Packages
     release o=Tailscale,n=bookworm,l=Tailscale,c=main,b=all
     origin pkgs.tailscale.com
 500 https://pkgs.tailscale.com/stable/debian bookworm/main amd64 Packages
     release o=Tailscale,n=bookworm,l=Tailscale,c=main,b=amd64
     origin pkgs.tailscale.com
 500 https://deb.nodesource.com/node_20.x nodistro/main amd64 Packages
     release o=. nodistro,a=nodistro,n=nodistro,l=. nodistro,c=main,b=amd64
     origin deb.nodesource.com

Look at the line starting with release and search for a combination that uniquely identifies the Caddy repository.
The output above is using the short form keywords, while the examples in 50unattended-upgrades use the long form. It's fine to use either.
One special case is the site keyword. This is the URL coming after origin in the output above and might be confusing.

Keywords

//   a,archive,suite (eg, "stable")
//   c,component     (eg, "main", "contrib", "non-free")
//   l,label         (eg, "Debian", "Debian-Security")
//   o,origin        (eg, "Debian", "Unofficial Multimedia Packages")
//   n,codename      (eg, "jessie", "jessie-updates")
//     site          (eg, "http.debian.net")

Based on the apt policy output above, here's what I use to enable automatic updates for these repositories.
Using origin and codename follows the standard Debian repos and I'd recommend using that if possible.
Node doesn't provide a reasonable repo file, so I had to set site based on the URL behind origin in apt policy

"site=deb.nodesource.com"; //Nodesource repository
"origin=Zabbix,codename=${distro_codename}"; //Zabbix Agent repository
"origin=Tailscale,codename=${distro_codename}"; //Tailscale repository

Switched after Louis' video about it. Haven't tried the voice yet, but just having swipe typing and suggestions again is a definite improvement.

Yeah, but you need root anyways to mount disks (most of the time), so doing a quick chown isn't that much effort.

Edit: chown > chmod

I use multiple subdirectories under /mnt for my fstab/systemd-mount managed disks. That includes local and network locations.

Download more RAM

Without docker you still just copy your files from Windows to Linux, though you have to find the right directories for that. Jellyfin can be installed directly on Debian. Just add their repo and go

You can't really install packages or modify configs on the host without root. Containers can only do some parts.

Maybe, but now I still need to remember the alias or distribute it to any machine I'm working on.
Not that difficult if you have everything managed with Ansible or similar anyways, but lots of people likely don't have that setup.

As someone who speaks a language with gendered pronouns but no neutral option, this is very awkward to deal with.

Never used Shopify unfortunately, so I can't help you with that.

The way I tag media is using MediaElch. It requires manually going through each series and identifying it, but with your proper naming it should give decent suggestions already.
If some metadata is missing for single episodes, try changing the metadata provider, sometimes one or the other just has bad/incomplete data.

As someone who writes bash scripts, fuck no, this is a terrible language and it shouldn't be used for anything more complex than sticking two programs together.

Also, parallelism goes right out of the window.

Maybe you'd convince me with a real programming language.