Deployment

Rates API deploys two Cloudflare Workers:

Worker	Serves
ratesapi-nz	API routes such as /api/v1/* and /openapi
ratesapi-nz-docs	The Fumadocs site and docs-only routes

Cloudflare D1 stores latest datasets and historical snapshots for the API Worker.

Prerequisites

1. A Cloudflare account
2. The project set up for local development
3. Bun installed
4. Wrangler CLI for Cloudflare Workers and D1 management

Cloudflare D1

Create the D1 database and apply the schema before deploying the API Worker.

wrangler login
wrangler d1 execute ratesapi-data --file=./apps/api/schema.sql

Worker Configuration

The API Worker is configured in apps/api/wrangler.toml. It owns the public API path routes on the apex domain:

[[routes]]
pattern = "ratesapi.nz/api/v1/*"
zone_name = "ratesapi.nz"

[[routes]]
pattern = "ratesapi.nz/openapi"
zone_name = "ratesapi.nz"

The docs Worker is configured in apps/docs/wrangler.toml. It owns the apex Custom Domain:

[[routes]]
pattern = "ratesapi.nz"
custom_domain = true

This lets the docs app serve https://ratesapi.nz/, while the API Worker still handles https://ratesapi.nz/api/v1/* and OpenAPI routes.

Dry-Run Deployment

Run the same deployability checks used by CI:

bun run deploy:dry-run
bun run build:docs
bun run deploy:docs:dry-run

Deploy

Deploy the API Worker:

bun run deploy

Build and deploy the docs Worker:

bun run build:docs
bun run deploy:docs

Post-Deployment Setup

• Confirm https://ratesapi.nz/ serves the docs app.
• Confirm https://ratesapi.nz/api/v1/health serves the API Worker.
• Confirm https://ratesapi.nz/openapi loads the Scalar API client.
• Confirm https://ratesapi.nz/openapi/json returns the generated OpenAPI document.
• Confirm the hourly scraping workflow has the CLOUDFLARE_API_TOKEN, CLOUDFLARE_ACCOUNT_ID, and D1 database environment it needs.

Production Considerations

• Keep route ownership clear: docs Worker owns the apex Custom Domain, API Worker owns API path routes.
• Review the deploy dry-run job when Worker bindings or route patterns change.
• Keep endpoint docs and OpenAPI schemas aligned with route changes.
• Monitor scraper failures separately from Worker uptime.