start.sh calls this when it can't fan out (no destinations, or none with a decryptable URL) instead of exiting straight into a silent runOnReadyRestart loop -- the 2026-07-28 no_dests incident was 83 relaunches in ~7 minutes with zero signal anywhere. Always logged and always recorded (for the dashboard's fanout_issue field, see api_stream_status); admin is alerted once per cooldown once the failure looks sustained rather than a one-off race at stream start.

The path is never logged (it carries the stream key) -- only the same one-way hash start.sh already prints as hash=..., so the two logs correlate without exposing anything.