> the worst that can happen is the company is down for a few hours overnight
I think this is a great example of why disagreements arise on HN: different world experiences and base assumptions. For many companies, being unavailable for that window of time would be catastrophic. We had one client that suffered about an hour of downtime (turned out to be their issue). They accounted that hour for 5 million dollars lost.
$5M/hour can pay for a lot of engineers. So as you say - with such assumptions - you can and should pay for both up-front design, QA and people on-call, otherwise you only have your self to blame for the loss.
I think this is a great example of why disagreements arise on HN: different world experiences and base assumptions. For many companies, being unavailable for that window of time would be catastrophic. We had one client that suffered about an hour of downtime (turned out to be their issue). They accounted that hour for 5 million dollars lost.