Deploying to Production

After the hard work of building your application, nothing is more exciting than deploying it to production. This guide explains how ZenStack participates in the deployment process.

Running ZenStack CLI During Deployment

If you use a framework that requires a build step, for example, "next build" for Next.js, you can run zenstack generate as the very first step in the build process. It ensures the schema.prisma file is up-to-date, and the supporting modules are generated into the node_modules/.zenstack directory.

Here's what the build script of our sample Todo app looks like for deploying to Vercel:

package.json

{
  "scripts": {
    "vercel-build": "zenstack generate && next build && prisma migrate deploy"
  }
}

Not Using ZenStack CLI During Deployment

There might be cases where you don't want to run the zenstack CLI during the deployment process. For example, you may want to avoid installing dev dependencies during the build to keep a clean "node_modules" folder. In such cases, you can run zenstack generate with a --output option to generate the supporting modules into your source tree and then commit them to your repository.

npx zenstack generate --output ./zenstack

At the runtime, when calling enhance(), you need to pass in an extra loadPath option to point to the directory where the supporting modules are located:

const db = enhance(prisma, { user }, { loadPath: './zenstack' })

In this way, all information ZenStack needs at runtime is already in the source tree, and you don't need to run zenstack generate during deployment.