Some applications, such as servers, often want to continue, even if a bug is detected in one part of the system, and the potential worst-case fallout from that specific kind of bug is limited. Though such bugs are often also logged, at the very least.

Different situations, different requirements, different options, different trade-offs.

Consider also the use cases in https://github.com/rust-lang/rust/issues/43596 , despite Rust defaulting to panicking. Also, despite Rust not having exceptions, some configurations allow for "catching" panics, meaning that some Rust programs have to be "unwind-safe", similar to C++ exception safety.