Building on Cloudflare

A form needs somewhere to save its data. An upload needs somewhere to live. Then requests arrive together, or a query takes longer than expected, and the choice of service starts to matter.

These guides follow decisions I’ve met while building two sites on Cloudflare. Each starts with a job the application needs to do, then works through the setup, the failure cases, and the cost. Start with the opening installment for a tour of Cloudflare’s ecosystem, how its services fit together, and where the five-dollar Workers plan fits.

Planned guides

The weekly series will cover these questions. Related products belong together when comparing them helps make the decision.

• How do you connect a domain, configure DNS, and check where requests go?
• What runs inside a Worker, and which limits affect the code you can put there?
• How do you serve a frontend, and when should you use Workers or Pages?
• Which responses should you cache, and when do tiered cache and Cache Reserve help?
• When does D1 fit, and how does Hyperdrive connect an existing Postgres database?
• When is KV enough, and when do updates need a Durable Object?
• Where should uploads live, and when are Images or Stream worth adding to R2?
• How do you send application email and route incoming messages?
• How do you protect an admin dashboard with Access and handle application sign-in?
• Which requests need Turnstile, WAF rules, bot controls, or rate limits?
• How do Workers AI, AI Gateway, Vectorize, and AI Search fit an AI feature?
• How do you build an agent or a remote MCP server with the Agents SDK?
• Which background work belongs in Queues, Workflows, Cron Triggers, or Pipelines?
• When do you need Containers, Sandboxes, or Browser Rendering?
• How does Tunnel connect services hosted elsewhere?
• What should you log and measure with traces, Analytics Engine, and Web Analytics?
• How do you deploy safely, inspect usage, and explain the monthly bill?