Guide & TutorialsYour first production deployment
Guide & Tutorials
Your first production deployment
Take a Northstar project from a local prototype to a monitored production release.
A reliable deployment starts before code reaches production. This tutorial prepares an environment, creates a release, and verifies that the application is healthy after traffic begins flowing.
Photo by Carlos Muza on Unsplash.
Prepare the environment
Create a dedicated production environment instead of reusing staging. This keeps credentials, variables, deployment history, and access controls isolated from day-to-day testing.
ts
const environment = await northstar.environments.create({
projectId: 'prj_01J8YQ2N7F',
name: 'production',
region: 'eu-west',
});Add secrets through the Northstar dashboard or your secret manager. Avoid passing sensitive values in command arguments because shell history and CI logs may retain them.
Create the release
Deploy an immutable commit rather than a moving branch reference. An exact commit makes the release reproducible and gives your team an unambiguous rollback target.
ts
const deployment = await northstar.deployments.create({
projectId: 'prj_01J8YQ2N7F',
environment: 'production',
commit: '8c31f82',
});Northstar moves the deployment through
queued, building, releasing, and ready. Treat a successful API response as acceptance of the request, not proof that the release is already serving traffic.Verify the result
Wait for the
deployment.succeeded webhook, then run a small application-level smoke test. Check the public health endpoint, load a representative page, and confirm that background jobs can reach their dependencies.Tip
Record the deployment identifier with your application release metadata. It connects production errors to the exact build and logs in Northstar.
Roll back safely
If verification fails, deploy the last known-good commit as a new release. This preserves a complete audit trail and avoids mutating an existing deployment.