I need to find the link, but I saw an incredible presentation at Gig City Elixir last year where they’d been running a globally distributed, in memory, medical database with no downtime or data loss for upwards of 5 years. One of the coolest projects I ever saw.

But yes, you do have to determine state and recovery patterns. Depends entirely on the situation but things like making sure your data will survive a process crash is straightforward.