Interesting point of view. Indeed, you probably want things to not be available 100% of the time. This forces you to see how resilient things really are.

I didn't know about this project. This sounds interesting, smart use of mkosi to make an Incus tailored system.

It was around two years ago, but maybe a good idea to revisit it with the recent AWS outage?

And a bunch of tool to use with it... But you can indeed do a lot with just SSH. This post gives a few good ideas.

Want to scare yourself with what might happen when you completely let go of your infrastructure? Here is an aggregator for that.

Definitely over engineered but a good way to play with many interesting tools.

Nice experiment. When looking at the actual infrastructure used, the servers are indeed nicely decentralized. For the users the picture would be different though.

Or why you need to own at least some part of your infrastructure.

Definitely not as fashionable as the kubernetes craze. This gives very interesting properties that multi-tenant applications can't really provide. The article is nice as it lays out properly the pros and cons, helps make the choice depending on the context.

Looks like an interesting tool for infrastructure automation. It's all Python based which is an interesting departure from yaml files in that space. Could be a nice alternative to Ansible. I might take it out for a spin.

Very fascinating piece. This shows the underappreciated job of maintaining the subsea cables needed for the Internet to function and how extreme the conditions can be. Definitely a peculiar life for the folks in that trade... also shows the repairs are clearly underfunded and that not enough people are embracing this career. And now, add geopolitics to the mix, it should make you wonder how all of this work at all and for how long it'll keep working.

When you do the math, the cloud offerings look very expensive for most workload indeed.

Looks like an interesting and comprehensive reference to squeeze as much reliability as possible from a Raspberry Pi.

Interesting stats, not that easy to gather. This gives a good overview of where the fediverse instances are hosted though.

Half a rant but interesting... Why are people making popular solutions to problems they'll never have? Just because it's been released by Google?

OK, I admit this looks like a very cool product. This could turn interesting for private infrastructures. Trying to get the benefits of cloud approaches while keeping it under control.

Interesting take... Is it really practical? Until which size is it viable? What are we loosing by aggregating? Also makes me wonder about alert fatigue... It clearly raises extra questions.

Even the giants are slowly moving back from microservices. DHH has a very cruel way to point it out, still that's true. Let's hope people realize the mistake it was in term of complexity.

Very interesting to see that move to owned hardware... turns out that not only the invoice is smaller in their case but the performances are much better as well.

Nice overview of what it takes to increase your uptime. It can get expensive quickly. This is also a good reminder that it's not only about software, it's a lot about people and administrative constraints as well.