This would be considered a failure, or are you saying they don't need to?

I am saying that in my experience they get upset when the VM or container they provision blows up because it lacks enough resources or they do not place guardrails on their app and end up getting OOMKilled.

And that frustration makes sense in the context of the article. Devs don't care about any of that stuff because they're customer facing, it's a distraction from their primary responsibility.

It would be like asking an Amazon delivery drivers to care about oil changes and tire rotations. It's much easier to have a team of mechanics whose primary responsibility is enabling drivers to just drive and focus on delivering packages.

I think my favorite interaction with a dev around this was when I was explaining how his java program looked like a big juicy target for the OOM killer and it had killed it in order to keep the system working. His response was, "I don't care about the system, I care about my program!" And he understood the irony of that, but it was a good reminder that we have somewhat different views and priorities.

A developer not caring about the system is why the file explorer is painfully slow today, compared to 15 years back.

If a programmes doesn't care about the system, I already know he's shit at his job.

That is basically WinUI/WinAppSDK, the whole WinRT stack and related dev experience, where even plain .NET is faster.

However the team will advertise it as performance, due to being written in C++.

Pity it gets slowed down by COM reference counting all over the place, which cannot be optimised away, and the application identity sandbox.