Latest posts
-
Minimal Observability: The Five Alerts You Actually Need
Error rate, p95 latency, oldest unprocessed job, dead-letter count above zero, and a daily money reconciliation. Everything else is noise at this stage.
-
Auth in a Weekend: The Setup That Doesn’t Bite Later
Magic links, 15-minute access tokens, a sessions table with real revocation, and rotation with reuse detection. Four tables, no password field.
-
The Early-SaaS Stack Decision Checklist
Hosting, database, auth, payments, and email: the opinionated default for each layer, plus the specific condition that should make you deviate.
-
The One-Server Question: What to Run Yourself Before You Have Customers
A single small VM covers every pre-revenue workload; the real decisions are which pieces to hand to a managed provider and which signals justify splitting the stack.
-
Infrastructure You Can Defer Until You Have Paying Customers
Start with the smallest auth, logging, and deploy setup that lets you reach the first users, then add layers only after real usage exposes the gaps.
-
Preview Environments That Isolate Production Data
Configure preview deploys so they run against separate data stores and never inherit production credentials or live records.
-
How to Set Up CI/CD for Next.js with GitHub Actions and Vercel
Build a reliable CI/CD pipeline that runs tests and type checks on every push before handing off to Vercel for preview and production deployments.
-
How to Set Up OpenTelemetry Tracing for Next.js with Grafana Cloud
Instrument a production Next.js app with OpenTelemetry, export traces to Grafana Cloud, and handle sampling, costs, and edge cases without vendor lock-in.
-
Lean Startup Infrastructure: What You Actually Need on Day One
The lean stack a startup genuinely needs — auth, observability, CI/CD, cost caps, security — and what you can safely ignore.