How do Kotlin systems behave under cascading coroutine failures?

Failures propagate through coroutine hierarchies unless explicitly handled with supervision.

By default, coroutine failure propagates upward and cancels sibling coroutines. This can cause cascading failure where one exception shuts down entire coroutine scopes. SupervisorJob prevents this by isolating failures. Without supervision, a single failure in a child coroutine can bring down entire workflows.