- `DecodedSample` → refresh groups → one ingestion run per group and window once the watermark passes → storage. - Port backend `precompute.rs`, `per_sample.rs`, series interner, one per PR. Tumbling windows only. - Then transfer relevant open ASAPQuery-backend issues and archive that repo. **Done when:** default plans (ingestion-time views) answer on the scoreboard. Part of #918. PR conventions: see #918 "PR stack".
Activity