I don't know why it is convenient to ignore the "managed" aspect of it.
DO's or AWS's managed databases do more than just stick it on an server VM. It has a firewall built it, replicas, backups, logs, API access to manage it, etc.
The reason why these managed database services even exist is because people are willing to pay for it. It provides them value. If it didn't, they'd stop working on them.
Processing web requests & some reports, pretty much the same as everyone do nowdays. Dev team add some features, but still on scale it the same - web requests and reports.
CPU is pretty much idling, like ~ 10-15% used on average, in spikes may be 20%. But that's minimal CPU hoster gives with RAM 128GB+ platform, so using it.
RAM (~ 110GB) obviously used for DB and the for auxilary tasks.
DB itself is medium sized, with raw data ~ 1.5TB yet
DISK IO on average is pretty high, around 55%, with often spikes to 100%, that part I'd add more IOPS on if response times grow above the limits.
Not saying you should do the same, just was a bit shocked and decided to share.