Documentation
Languages › Node.js

CI/CD in Node.js with Dagger

Run Node.js CI/CD in Dagger with an encrypted .env.vault file

Find a complete code example on GitHub for this guide.

Initial setup

Install the Dagger Node SDK.

npm install @dagger.io/dagger --save-dev

Create a ci/index.mjs file.

ci/index.mjs

// ci/index.mjs
import { connect } from "@dagger.io/dagger"

connect(async (client) => {
  const node = client.container().from("node:16-slim").withExec(["node", "-v"])

  const version = await node.stdout()

  // print output
  console.log(`Hello ${process.env.HELLO}.`)
}, { LogOutput: process.stdout })

Run the pipeline.

node ci/index.mjs

8: [0.11s] v16.20.1
8: exec docker-entrypoint.sh node -v DONE
Hello undefined.

Docker is required to be running on your machine for Dagger to work.

Once run, the pipeline build will say 'Hello undefined.' as it doesn't have a way to access the environment variable yet. Let's do that next.

Install dotenv

Install dotenv.

npm install dotenv --save-dev # Requires dotenv >= 16.1.0

Create a .env file in the root of your project.

.env

# .env
HELLO="World"

As early as possible in your ci pipeline, import and configure dotenv.

ci/index.mjs

// ci/index.mjs
import 'dotenv/config'
console.log(process.env) // remove this after you've confirmed it is working

import { connect } from "@dagger.io/dagger"

connect(async (client) => {
  const node = client.container().from("node:16-slim").withExec(["node", "-v"])

  const version = await node.stdout()

  // print output
  console.log(`Hello ${process.env.HELLO}.`)
}, { LogOutput: process.stdout })

Try running it locally.

node ci/index.mjs
{
  ...
  HELLO: 'World'
}
Hello World.

Perfect. process.env now has the keys and values you defined in your .env file.

That covers local simulation of the CI. Let's solve for the real CI environment next.

Build .env.vault

Push your latest .env file changes and edit your CI secrets. Learn more about syncing

npx dotenv-vault@latest push
npx dotenv-vault@latest open ci

Use the UI to configure those secrets per environment.

dotenv.org

Then build your encrypted .env.vault file.

npx dotenv-vault@latest build

Its contents should look something like this.

.env.vault

#/-------------------.env.vault---------------------/
#/         cloud-agnostic vaulting standard         /
#/   [how it works](https://dotenv.org/env-vault)   /
#/--------------------------------------------------/

# development
DOTENV_VAULT_DEVELOPMENT="/HqNgQWsf6Oh6XB9pI/CGkdgCe6d4/vWZHgP50RRoDTzkzPQk/xOaQs="
DOTENV_VAULT_DEVELOPMENT_VERSION=2

# ci
DOTENV_VAULT_CI="x26PuIKQ/xZ5eKrYomKngM+dO/9v1vxhwslE/zjHdg3l+H6q6PheB5GVDVIbZg=="
DOTENV_VAULT_CI_VERSION=2

Set DOTENV_KEY

Fetch your CI DOTENV_KEY.

npx dotenv-vault@latest keys ci
# outputs: dotenv://:[email protected]/vault/.env.vault?environment=ci

Set DOTENV_KEY for Dagger.

DOTENV_KEY='dotenv://:[email protected]/vault/.env.vault?environment=ci' node ci/index.mjs

The result will be:

8: exec docker-entrypoint.sh node -v CACHED
Hello ci.

That's it! Your .env.vault file is decrypted and its CI secrets injected as environment variables – just in time.

You'll know things worked correctly when you see 'Loading env from encrypted .env.vault' in your logs. If a DOTENV_KEY is not set (for example when developing on your local machine) it will fall back to standard dotenv functionality.

You succesfully used the new .env.vault standard to encrypt and deploy your secrets. This is much safer than scattering your secrets across multiple third-party platforms and tools. Whenever you need to add or change a secret, just rebuild your .env.vault file and redeploy.