dag-deploy

Submit a Helm chart to the DAG server and monitor the deployment.

Usage

dag-deploy <chart-path> [options]

Arguments

Argument	Description
<chart-path>	Path to a Helm chart directory, .tgz, or .tar.gz file

If a directory is provided, it will be packaged into a .tgz tarball automatically.

Options

Option	Env Variable	Default	Description
--server <url>	DAG_SERVER_URL	—	DAG server URL (required)
--repo <url>	DAG_REPO_URL	auto-detected	Override the repository URL
--job-id <id>	DAG_JOB_ID	auto-detected	Override the CI job ID
--job-token <token>	DAG_JOB_TOKEN	auto-detected	Override the CI job token
--deploy-version <version>	DAG_DEPLOY_VERSION	—	Deployment version (required)
--timeout <seconds>	DAG_TIMEOUT	300	Client-side timeout for waiting on deployment status (seconds)
--values-file <path>	—	—	YAML file to deep-merge into the chart's base values.yaml
--set <key=value>	—	—	Set a dotted path to a literal string value (repeatable)
--set-json <key=json>	—	—	Set a dotted path to a JSON-parsed value (repeatable)
--set-file <key=filepath>	—	—	Set a dotted path to the contents of a file (repeatable)

CI Auto-Detection

When running in a CI environment, dag-deploy automatically detects the repo URL, job ID, and job token from environment variables:

GitLab CI (detected via GITLAB_CI):

Value	Source
Repo URL	CI_PROJECT_URL
Job ID	CI_JOB_ID
Job Token	CI_JOB_TOKEN

GitHub Actions (detected via GITHUB_ACTIONS):

WARNING

Server-side GitHub job token verification is not yet implemented. GitHub Actions auto-detection works on the CLI side, but deployments will fail at the server validation step.

Value	Source
Repo URL	GITHUB_SERVER_URL/GITHUB_REPOSITORY
Job ID	GITHUB_RUN_ID
Job Token	GITHUB_TOKEN or ACTIONS_RUNTIME_TOKEN

Exit Codes

Code	Meaning
0	Deployment succeeded (deployed status)
1	Deployment failed (failed status or error)

Values Overrides

dag-deploy can modify the chart's values.yaml before packaging and uploading. When --values-file is provided, the specified YAML file is deep-merged into the chart's base values.yaml — nested objects are merged recursively, while scalars and arrays from the overlay replace the base values. Then --set, --set-json, and --set-file overrides are applied on top.

Values overrides require the chart path to be a directory (not a .tgz).

dag-deploy ./charts/my-app \
    --server https://dag.example.com \
    --deploy-version v1.2.3 \
    --values-file values-production.yaml \
    --set image.tag=v1.2.3

TIP

Values set via --set are always strings. Use --set-json for typed values (numbers, booleans, objects, arrays):

--set-json 'resources={"limits":{"cpu":"500m","memory":"256Mi"}}'
--set-json 'replicas=3'

Examples

Deploy a chart directory:

dag-deploy ./charts/my-app \
    --server https://dag.example.com \
    --deploy-version v1.2.3

Deploy with values overrides:

dag-deploy ./charts/my-app \
    --server https://dag.example.com \
    --deploy-version v1.2.3 \
    --values-file values-staging.yaml \
    --set image.tag=$CI_COMMIT_SHA \
    --set-json 'replicas=3' \
    --set-file config.data=./config.json

Deploy a pre-packaged tarball:

dag-deploy my-app-1.2.3.tgz \
    --server https://dag.example.com \
    --deploy-version v1.2.3

Deploy with explicit credentials (outside CI):

dag-deploy ./chart \
    --server https://dag.example.com \
    --deploy-version v1.2.3 \
    --repo https://gitlab.com/org/my-app \
    --job-id 12345 \
    --job-token glcbt-abc123