Last 1 Infrastructure Tier Resequencing (2026-08-28)¶
Decision owner: Ryan
Cursor status: Planning only. No autonomous tier kickoff until Ryan picks first product.
Strategic reframe¶
| Old assumption | New scope |
|---|---|
| Tier 3 = admin compute-only; Supabase stays hosted until Tier 6 | Every product: migrate Railway compute + Supabase database to R940 |
| Admin early (Tier 3) | Admin last — downstream BFF reconnects once to sovereign upstreams |
| Partial sovereignty narrative | Full sovereignty story: no hosted Postgres for Last 1 apps when complete |
New tier order (proposed)¶
| Tier | Product | Scope | Notes |
|---|---|---|---|
| Done | last1.id | Full stack | auth.last1.id on R940 (Tier 2 closed 2026-08-27) |
| Done | n8n | Compute | Tier 1 |
| Done | LastVet API | AWS/R940 clinical | Separate from this track |
| 3 | RealOutcomes (+ IRS cron) | Railway + Supabase → R940 | Authorized 2026-08-28 — T3-0 audit complete |
| 4 | Onset clinic + Supabase | Full stack | After Tier 3 |
| 5 | last1.app + RadioCheck bundle | Shared Supabase tcthzakppndkykatoefx |
After Tier 4 |
| 6 | last1-admin | Reconnect BFF to sovereign upstreams + cutover + teardown | Resumes paused VM 115 work |
Admin is no longer Tier 3. Prior T3-1…T3-3 work is preserved on VM 115; execution paused (TIER_3_ADMIN_PAUSED_STATUS.md).
First product candidates (Ryan decision pending)¶
Option A — RealOutcomes (Ryan leaning)¶
| Pros | Cons |
|---|---|
| Pre-user / lower blast radius | Stripe, edge functions, IRS cron coordination |
Own Supabase project (cpocwrygxlqhzvmlrpjv) — clean boundary |
Marketing/app split at apex |
| Strong full-stack rehearsal before clinical | GitGuardian D-live rotations (Twilio/Stripe/Claude) natural fit |
| Good SBIR "first full product on R940" anchor |
Muster item: migrate-realoutcomes-railway-supabase-to-r940
Option B — Onset clinic¶
| Pros | Cons |
|---|---|
| Real users; SOC 2 driver | PHI-adjacent — highest care |
| Separate Onset Supabase | Mobile client env cutover |
| admin already proxies Onset dashboards |
Muster item: migrate-onset-clinic-supabase-to-r940
Option C — last1.app + RadioCheck bundle¶
| Pros | Cons |
|---|---|
| Largest sovereignty win when done | Shared Supabase tcthzakppndkykatoefx — must move as one cutover |
| Unblocks shared auth DB story | Edge Functions, Realtime, multi-app coupling |
| Tier 6 in old plan | Biggest scope — poor first rehearsal |
Muster item: migrate-last1app-radiocheck-shared-supabase-to-r940
What Cursor waits for¶
- ~~Ryan names first product~~ Done: RealOutcomes (Tier 3).
- Ryan reviews T3-0 audit (TIER_3_REALOUTCOMES_MIGRATION_AUDIT.md).
- Ryan authorizes T3-1 planning (migration plan, Edge Function strategy, VMIDs, BWS prefix).
- No provisioning/dump/deploy until T3-1 sign-off.
Paused work (do not discard)¶
- VM 115 + BWS + soak + tunnel prep: see TIER_3_ADMIN_PAUSED_STATUS.md
- T3-4 test plan exists but not executed until admin tier resumes
References¶
- TIER_3_ADMIN_MIGRATION_AUDIT.md (T3-0; compute-only framing superseded by this doc)
- Muster:
migrate-last1-admin-railway-to-r940(hold), product migration items under engineering/hq