[toast0]

I think you're basically right. At the end of the day, if the servers catch fire, you probably need to call the responsible developers for help getting things back on track.

The real questions then are how is the work divided, what does escalation look like, what steps are taken to reduce incidents outside of office hours, what's being done to allow for urgent maintenance to wait until office hours, etc, and also like how well is that all working, and where is the leadership in this?

I worked somewhere with not great answers to a lot of these questions, but it was okish because the cofounders were waking up with all the pages and hoping on to fix things too, and I had a mediocre experience with a dedicated ops team at a previous job that I didn't want to repeat. Eventually, as the team got bigger, we made a lot of things better, including on call responsibilities.