reliably maintaining services

it’s funny that you use that as a selling point.

In my experience almost no outage happens because hardware failures. And most outages happen because bad configurations and/or expired certs, which in turn are a symptom of too much complexity.