Ship & operate / 5 min read
Your first production deployment
A practical checklist for taking a project from localhost to a public URL, with reliable builds, persistent data, and a recovery plan.
Make the build repeatable
Pin your runtime and commit your dependency lockfile. Run the production build locally and document the start command. Keep secrets out of the repository and configure them in the host's environment settings. Use separate production and development credentials.
Treat data as persistent infrastructure
Deployments may replace the filesystem. Put uploads on persistent storage and choose a database that survives redeploys. Plan migrations, create backups, and test restoration in a separate environment. Never assume a successful backup means a successful restore.
Connect the domain safely
Use the DNS records supplied by your hosting provider. Check existing mail records before changing DNS, allow for propagation, and verify HTTPS on both your preferred hostname and its redirect. Do not remove records you do not understand.
Check the real production path
Test navigation, forms, error pages, mobile layouts, and login in the deployed environment. Verify canonical URLs and ensure private administration routes are not indexed. Use a non-sensitive test account; do not put real customer data into test fixtures.
Know how to recover
Set up health checks and actionable logs. Document how to roll back a release, restore data, rotate credentials, and contact the provider. Measure latency and errors before increasing capacity so that you solve the actual bottleneck.