feat(roster): attribute standalone endpoint-list waits by stage

Add opt-in, thread-scoped ROSTER_WAIT records over the complete-roster path so roster reads, registry open/tune/schema/query work, the UNLISTED worklist and each presence probe stage are separated rather than summed. Every record names its budget type and producer progress on success and failure, and an outer observation cutoff is recorded as a cutoff rather than producer completion. Output bytes, roster rows and existing deadlines are unchanged. Source-only under the execution hold; no run has been measured with it.

Co-authored by: todlando