Staging deploy & public status (OPS-4)

Staging Worker

Template: apps/worker/wrangler.staging.toml

1. Create staging D1, KV, R2 with unique names.
2. Replace REPLACE_WITH_* IDs in the staging toml.
3. Deploy:

cd apps/worker
pnpm exec wrangler deploy src/worker.js --config wrangler.staging.toml

4. Point dashboard staging env at the staging Worker:

NEXT_PUBLIC_FLUXYCHAT_WORKER_URL=https://<staging-worker-host>

GitHub Actions: workflow Deploy → environment staging (requires CLOUDFLARE_API_TOKEN, CLOUDFLARE_ACCOUNT_ID).

Public status page

Dashboard route: /status (server-rendered, polls GET /health on the configured Worker).

Optional custom domain:

1. Add status.fluxychat.com (or your domain) in Vercel/hosting DNS → dashboard project.
2. Link from marketing footer and console (already wired to /status).

Health payload fields: ok, degraded, checks, platformBindings, paymentsEnabled.

Console operational alerts (RD-6)

Signed-in admins with an active project see a banner when GET /stats/alerts reports openAlerts > 0. Link goes to Analytics.

CI gate before deploy

.github/workflows/deploy.yml runs pricing/tenant/openapi/DO checks, worker typecheck, and pnpm test before wrangler deploy.