Treat post-disconnect SIGKILL as transport teardown failure

When a LiveKit or Pipecat per-call process receives SIGKILL shortly after its room is deleted, first inspect teardown diagnostics rather than blaming the container platform or an event-loop freeze. A dangling transport audio-consumer task can survive call recording and database finalization, keeping the interpreter alive until the parent’s bounded shutdown deadline expires. Fix ownership and cancellation of the transport task; keep the supervisor deadline as containment, not the normal exit path. Confirm with stable container identity, no OOM/restart, room deletion preceding SIGKILL, and the same dangling task on affected workers.

Applies when: isolated voice workers are force-killed only during post-disconnect cleanup. Keywords: LiveKit Pipecat SIGKILL dangling audio task teardown room deletion worker cleanup