Repository navigation
Description
🤖 Generated by the Daily AI Assistant
Problem
The wedding-app has all core features (login, RSVP, room booking, program, gallery, admin view) and is production-deployed. The wedding is May 16, 2027. The product direction shifted from feature-delivery to polish, reliability, and event operations: ensuring the app holds up under real guest traffic, that the couple can act on RSVP data efficiently, and that the guest experience is accessible and fast on any device.
Phase 1 — MVP → wedding-ready ✅ DELIVERED
Three themes closed the gap from MVP to wedding-ready:
- Guest experience quality — accessibility and performance.
- Administrative tools — the couple can act on RSVP data.
- Ops reliability — lightweight quality gates to catch regressions before guests arrive.
Children
- feat: axe-core a11y guard in Playwright CI — feat: axe-core a11y guard in Playwright CI #92
- feat: admin RSVP export to CSV — feat: admin RSVP export to CSV #93
- feat: RSVP deadline enforcement — feat: RSVP deadline enforcement #94
- perf: Lighthouse CI performance budget — perf: Lighthouse CI performance budget #95
All four shipped. The admin view also gained an at-a-glance summary (attending / declined / pending / bookings) and per-pair dietary + booking detail. The app is HA (3-instance CloudNativePG, 2 app replicas + PDB + topology spread), backed up daily to R2 with PITR, and hardened (restricted securityContext, NetworkPolicy, External-Secrets-managed admin code).
Phase 2 — operations & polish (post-MVP)
Strategy-review assessment (2026-06-22): the app is mature and well-built; with ~11 months to the wedding and only 14 guest pairs, Phase 2 is intentionally small — targeted ops-reliability and optional UX polish, not new feature surface. Over-scoping here would be manufactured work.
Children / candidates
- ops: verify wedding-db backup restore — ops: verify wedding-db backup restore (recovery drill for irreplaceable RSVP data) #133 (the RSVP/booking data is irreplaceable and backups are configured but the restore path is untested — a recovery drill + runbook proves it works; ready)
- perf: full Lighthouse CI Core Web Vitals budget — perf: full Lighthouse CI Core Web Vitals budget (deferred from #95) #100 (deferred from perf: Lighthouse CI performance budget #95; ready)
- (candidate, maintainer-decision) Guest RSVP confirmation / receipt — today an RSVP POST persists silently with no guest-facing confirmation. ROI is marginal at 14 pairs (the couple may prefer to confirm personally); file only if wanted.
- (candidate, maintainer-decision) Invitation-window alerting — a lightweight uptime / RSVP-write-error alert so an outage during the critical RSVP period is noticed. Likely best solved at the platform observability layer (coroot) rather than in-app.
Rough size
Phase 2 is ~1–2 small ready items (#133, #100) plus two optional candidates the couple may choose to skip. Suggested order: #133 → #100.
Metadata
Metadata
Assignees
Labels
Type
Fields
Projects
- StatusShow more project fields🧊 Icebox
Activity
devantler commentedon Jun 10, 2026
Implementation order & notes
Suggested order: #92 → #93 → #94 → #95
#92 (a11y guard): Start here — zero CI changes needed (runs in existing
npm run test:e2e). Fixes any existing violations before the other changes add new routes.#93 (RSVP export): Simple server-only CSV endpoint. Prerequisite: a real DB connection for manual verification, but unit tests work with the mock layer.
#94 (RSVP deadline): Build after #93 so the admin view already shows the closed state clearly. The deadline env var pattern is additive — safe to merge independently.
#95 (Lighthouse budget): Last, since it depends on the final build state and is most likely to need threshold tuning. If the Node adapter + Lighthouse port-binding proves unreliable in CI, fall back to
bundlesizeand track as a follow-up.What was verified NOT needed:
admin-code-externalsecret.yaml+ platformopenbaoSecretStore; no SOPS)copilot-instructions.md→ already present in.github/ci.yamlstructuredevantler commentedon Oct 17, 2026
Strategy review, 2026-10-17
Where the app is. Feature-complete and in the polish and reliability phase this epic describes. Every child linked here is closed, and so are both Phase 2 candidates in the body (the backup restore drill #133 and the Core Web Vitals budget #100). Recent work hardened delivery: one declared Node toolchain (#388), release of runtime and base-image updates (#408), and two RSVP and admin fixes today (#410, #411).
Delivery is still the weak link. The latest release is v1.15.19 (2026-10-11). #408 and today's fixes merged after it, and the release runs for them are queued behind saturated Actions runners, not failing. #407 stays open until a release that includes the SvelteKit 3 move (#402) is published; check that on the next run rather than re-opening the fix. Two gaps remain beyond #408:
Roadmap, in the order I recommend:
Nothing here needs new guest-facing features before the wedding; the remaining work keeps what guests use current and reliable.
Upstream release review (key dependencies):
build(deps-dev)and will not release on their own; that is #422.Decision for the maintainer: none required. If you want this epic to mirror the remaining work more tightly, its body could gain a Phase 3 list with the items above.