System iNEWS magazine recently published the "Top 10 IT Problems We Don't Have with IBM i". I don't want to paste the whole thing here and I don't want to put the link because it needs a subscription and people who don't know IBM i really shouldn't bother subscribing. So I'll just mention two somewhat-related items on the list.
#2 Unplanned Downtime
#7 Rebooting as a "Solution"
For four years, I worked for a company that ran AS/400 (a.k.a. iSeries a.k.a System i) along side another system, which I won't mention. And it got the reputation of the most "non-problematic" system. I mean the box itself was a dwarf compared to the racks and racks of servers needed to run the other system, and yet it's very stable. And it's not stable because it's rarely used. In fact, it's used 24 hours a day, 6 days a week, and up 24/7. The other system had, in total, more users than the AS/400, but in context, the lone AS/400 (with just one CPU) handled more users per CPU compared to the other system. The only time that I could remember it was taken down was because of a problem with the power supply of the data center where it was located and then for system upgrade. In the same span, the other system was taken down unplanned several times because of, among other things, problem with the database (several times), and problem with something that I didn't entirely understand (it refused to let anyone sign in). Of course all these are just based on observation, as a developer for both systems. Just anecdotes and aren't scientific. I didn't really know the root causes of what happened. The AS/400 just appeared stable. It wasn't even part of the yearly planned downtime (during long Holy Week weekend of the year) per se because for some reason it wasn't needed, though it's still affected because the network itself was taken down for maintenance.