Say you panic, now you need to have an external system that catches this panic and reports back; and does something meanwhile to recover your system.

If you think about it, it’s not really different from handling the bubbled up error inside of Rust. You don’t (?) your results and your errors go away, they just move up the chain.