I recommend Matrix with the Conduit server. This server requires almost no resources and even runs on a Raspberry Pi.

Cinny works perfectly as a desktop client (in case you want to escape from the ubiquitous Element). And for mobile I would use Element for Android/iOS although FluffyChat also works very well.

It is not about leading anything but about having the code in a repository so that it is easy to read/consult/audit/etc.

You can upload the code to any service (it doesn't have to be GitHub, it can be GitLab, sourcehut, etc...) and disable issues and comments.

IMHO, if the protocol specifications seem like pedantry to you the best thing to do is either drop it or design your own protocol. But if in the end what you do is to release a product that is only half-compliant, the rest of the community will turn their backs on you because, by not complying with the standard, you can produce all kinds of unexpected situations.

Before using borg I would recommend you to take a look at restic. In my opinion it is better in everything than borg.

As for how to backup the database, my advice is to export the database to a SQL file and backup that file. That will always be easier than having to deal with agents that connect to the database.

As for the number of repositories, if you use restic, a single repository is enough. Besides, as restic does deduplication, if you have the same files between your machines, they will only occupy the space of one. ;-)

I hope I have helped you with some of my ideas.

Best regards.

If you have an NVIDIA card don't upgrade before see here, here and here.

If you live in Europe you have 1TB by € 3.81 / month with Hetzner. It works fantastic with Restic (I'm using it too for my backups).

Apart from the registries you have in GitLab and GitHub if you are looking for something more generic like Docker Hub you have Quay (from RedHat). It works very well and has a pretty nice interface (especially the new one that is in testing).

You don't really need Cloudflare to have your own domain, you can do everything directly with GitHub.

Restic (or rustic) and Hetzner storage box 🤩

I'll tell you how I do it. That doesn't mean it's the best way or anything like that. First, you need to have plenty of unit tests (which the AI can also help with); that ensures that even if the AI makes changes, it has to do so under strict rules. Then, it's also best to have a technical specifications document where you outline what you want to do and under what conditions.

From there, the next step is to run a thorough set of running tests on the results generated by the AI (because it’s faster than reviewing the code itself). Then, once you have something consistent, you can start tweaking the code yourself.

Ultimately, what I always do is let the AI write the commit message. It’s infinitely better than me summarizing the change, and it’s even easier for me to read.

Best alternatives:

• GitLab
• sourcehut
• Codeberg
• Forgejo on Disroot

Self-hosted:

• Gogs
• Forgejo
• GitLab

Other alternatives:

• Bitbucket
• Launchpad
• SourceForge I do not recommend this one, but I put it anyway.

LessPass has the possibility to connect to a database (via its API) to store the configurations made for each site. This API can be used from any of the clients (either the browser extension, the mobile application, etc.).

You set up the DB server wherever you want. If you want something light you can use this implementation. And if you are interested, there is also a command line client.

Running any AUR helper (I recommend paru instead of yay but either is a good option) should always be done with a normal user and never with sudo.

This is because both the download of PKGBUILD as well as the download of the sources and the compilation must be done with a normal user. Doing that as root poses a HUGE risk to the integrity of the system, a bad PKGBUILD, a source code with errors, a script with a space where it shouldn't be and you can even break your system completely, you could even execute malicious code!

It is the AUR helpers themselves that will (eventually) ask you for a sudo authentication for when they need to do the final step of installing the package once it has been built.

restic without any doubt. I use it with S3 backend and SSH copy and it has an excellent performance (with copies of years).

Borg I was using it for a while (to compare) and I do not recommend it, it is not a bad product, but it has a lousy performance compared to restic.

Kopia I didn't know it, but from what I have read about it it seems to be very similar to restic but with some additions to make it pretty (like having ui).

Some people say that Kopia is faster in sending data to the repository (and other people say it's restic), I think that, unless you need ui, I would use restic.

I know this is not the best answer since you would probably like me to talk about Nebula, but I have to say that the best solution I have found for setting up a mesh VPN is Zerotier.

It is a very complete solution. Multisystem, very simple but very configurable, fast, etc.

You simply start by creating a network on the public controller (which will generate an ID for that network) and then join the rest to that network and everyone can communicate with everyone (by default, then you can create subnets if you want).

Using the public controller is completely optional (I personally use it because it is convenient for me and because I have few hosts) but if you want you can set up your own controller, I have an article (the bad thing is that it is in Spanish, but if you run a translator you can understand it perfectly) where I explain how to do it without any requirement. If not, you can use ztncui for it.

Take a look at it, you might find it more attractive than Nebula.

By the way, for me one of the great advantages of ZeroTier is that I don't have to worry about certificates and keys, the controller takes care of everything for you and security is guaranteed from the point of view that each node has a unique identifier.

I was an Inoreader user too, but I discovered Miniflux and cannot be happier 😉

About that I have made the simplest possible implementation of a syslog server in Rust. I use it to collect OpenWrt logs and store them on my server. ;-)

Well, if you want private images it is normal that they charge you for it. What I advise you to do is to make the images public and mount the private part as a volume. This way you can upload the images wherever you want without worrying.

Another option if you want the resulting image to have something private is to create as much as you can in a public image and have a script that adds the private part as the last layer.

Yes. The first thing I did was create an extremely detailed specification document where I wrote down what I wanted, how I wanted it and where I wanted it. For me, that’s the most important foundation. With that, you’ve already done a lot of the work. In fact, I even go so far as to write pseudocode in some cases.

From there, you use that as your ‘bible’ and stick to the script you’ve written for yourself. Obviously, as you go along, unforeseen issues may arise, and in that case, make a note in the spec, find the best approach, and carry on.

Yes, without a doubt, for me it is the most balanced client, a pity that there is not for Android, but well, in mobile Element does not give problems either.