001 — Cloudflare Workers + D1 + Wrangler

Context

The prototype needs an HTTP API, scheduled DNS verification, SQLite-scale persistence, and a simple deploy path. Splitting “API elsewhere + managed Postgres + separate scheduler” would add moving parts without teaching the domain model.

Decision

Run the server on Cloudflare Workers with D1 for storage and Wrangler for local/dev and deploy (including the static web app). Accept D1/SQLite constraints (batches, limited concurrency, dialect specifics) as the cost of that simplicity.

Consequences

• Cron is Wrangler scheduled triggers; schema and migrations are D1-oriented.
• Local/dev uses wrangler and .dev.vars.
• Scaling beyond single-region SQLite or needing heavy relational features would force a later platform rethink — not a goal of this prototype.