Reliability often comes at the expense of making some people's lives much less pleasant (especially in situations of underfunding or not having the corporation back any of your SRE initiatives), or, alternatively, with exponentially increasing costs, since all of the sudden everything needs to be high availability and have redundancy in place.

Consider the recent Roblox outage. No one died because of it, yet countless engineers were working around the clock to ensure the profitable operation of the platform.

When the Latvian Electronic Tax Declaration system goes down on the first day of the year when you can declare your taxes (basically every year), the ministry representatives shrug their shoulders and essentially say: "Nah, we can't make it cope with the load. Just wait a day." And you know, maybe they can't, but at the same time even so everyone still manages to submit their taxes eventually.

Thus, perhaps it's worth it to consider when and where we need the scalability and high availability, e.g. government centralized auth services under Latvija.lv which also went down because of the load and as a consequence people had to wait longer to sign up for COVID vaccines and couldn't use other systems until it was resolved: https://blog.kronis.dev/articles/manavakcina-lv-and-apturico...

Then, we should focus on those more important cases. Now, I'm not trying to suggest that the incompetence or dismissive attitudes are (always) okay, instead, I'm merely suggesting that we should consider our priorities and wonder what a more laid back environment would look like.

I still recall once sending a hotfix for a prod issue at 3 AM. I wonder what things would look like if instead of worrying whether a certain business process can take place at a certain date, the attitude would have been more like: "We'll just tell everyone to wait a day or two. It's not like our product is time sensitive."