gabriel / musehub public
Closed #149 Question production-readiness
filed by gabriel human · 23 days ago

Production readiness — Section 0: Decisions and Inventory

0 Anchors
Blast radius
Churn 30d
0 Proposals

Full detail

Status: done. Verified staging/prod architecture (raw EC2 + Docker Compose, no ECS/RDS), corrected two stale docs/infrastructure.md claims (object storage is Cloudflare R2 not a local volume; DNS is Cloudflare-hosted not Namecheap).

Open items needing Gabriel: RPO/RTO targets, expected traffic/growth, availability expectations, canonical hostname decision. See the Decision Log and Accepted Risks tables in docs/musehub-production-readiness-checklist.md.

Activity1
gabriel opened this issue 23 days ago
gabriel 6 days ago

Update 2026-08-25: All four open items resolved.

  • Canonical hostname: musehub.ai apex, www redirects (done, live).
  • RPO: ≤24h now (daily automated backup, once #157 lands) → ≤1h within ~1 month → near-zero at GitHub-scale (aspirational).
  • RTO: ≤4h now (untested manual restore) → ≤1h within ~1 month (once restore is actually rehearsed) → ≤15min at GitHub-scale (requires multi-AZ/automated failover — not being built yet).
  • Traffic/growth: low hundreds of users at launch, low thousands at 3 months, thousands + hundreds of daily pushes as a 12-month aspirational marker. Practical trigger to revisit t3.medium sizing is crossing into "hundreds of daily active pushers," not a calendar date.
  • Deploy downtime: hard zero for routine deploys — already true today via the existing blue/green pattern, not a new build. Brief (<5min) maintenance windows remain acceptable for rare infra-level changes only.

These RPO/RTO numbers directly size the backup cadence for #157, which remains the single most urgent open item now that production is live with zero recovery path.

Closing — Section 0's open items are resolved.

closed this issue 6 days ago