CLI Commands Reference

Complete reference for all deployml CLI commands.

Core Commands

`deployml deploy`

Deploy infrastructure based on a YAML configuration file.

Usage:

# Basic usage
deployml deploy --config-path config.yaml

# Skip confirmation prompts
deployml deploy --config-path config.yaml --yes

# Generate manifests only (GKE)
deployml deploy --config-path config.yaml --generate-only

Options: - --config-path, -c: Path to YAML config file (required) - --yes, -y: Skip confirmation prompts - --generate-only, -g: Only generate manifests, do not apply (GKE)

Example:

# Deploy Cloud Run stack
deployml deploy -c cloud-run-config.yaml

# Deploy GKE stack with manifest generation only
deployml deploy -c gke-config.yaml --generate-only

`deployml destroy`

Destroy infrastructure and optionally clean up workspace.

Usage:

# Destroy infrastructure
deployml destroy --config-path config.yaml

# Clean workspace after destroy
deployml destroy --config-path config.yaml --clean-workspace

# Skip confirmation
deployml destroy --config-path config.yaml --yes

Options: - --config-path, -c: Path to YAML config file (required) - --workspace: Override workspace name from config - --clean-workspace: Remove entire workspace after destroy - --yes, -y: Skip confirmation prompts

Example:

# Destroy and clean workspace
deployml destroy -c config.yaml --clean-workspace --yes

`deployml init`

Initialize cloud project by enabling required APIs.

Usage:

# Initialize GCP project
deployml init --provider gcp --project-id YOUR_PROJECT_ID

# Initialize AWS project
deployml init --provider aws

Options: - --provider, -p: Cloud provider (gcp, aws, azure) - --project-id, -j: Project ID (for GCP)

Example:

deployml init -p gcp -j my-project-id

`deployml generate`

Generate a deployment configuration YAML file interactively.

Usage:

deployml generate

Example:

# Interactive config generation
deployml generate
# Follow prompts to create config.yaml

`deployml doctor`

Run system checks for required tools and authentication.

Usage:

# Basic check
deployml doctor

# Check GCP APIs
deployml doctor --project-id YOUR_PROJECT_ID

Options: - --project-id, -j: Project ID (for GCP API checks)

Example:

deployml doctor --project-id my-project-id

GKE Commands

`deployml gke-init`

Generate Kubernetes manifests for GKE.

Usage:

# Generate MLflow manifests
deployml gke-init \
  --output-dir ./manifests/mlflow \
  --image mlflow-demo:latest \
  --project YOUR_PROJECT_ID \
  --service mlflow

# Generate FastAPI manifests
deployml gke-init \
  --output-dir ./manifests/fastapi \
  --image fastapi-demo:latest \
  --project YOUR_PROJECT_ID \
  --service fastapi \
  --mlflow-uri http://mlflow-service:5000

Options: - --output-dir, -o: Directory to create Kubernetes manifests (required) - --image, -i: Docker image name (local or GCR) (required) - --project, -p: GCP project ID (required) - --service, -s: Service type: mlflow or fastapi (default: mlflow) - --mlflow-uri, -m: MLflow URI (for FastAPI)

Example:

deployml gke-init \
  -o ./manifests/mlflow \
  -i mlflow-demo:latest \
  -p my-project-id \
  -s mlflow

`deployml gke-deploy`

Deploy Kubernetes manifests to GKE cluster.

Usage:

deployml gke-deploy \
  --manifest-dir ./manifests/mlflow \
  --cluster my-gke-cluster \
  --project YOUR_PROJECT_ID \
  --zone us-west1-a

Options: - --manifest-dir, -d: Directory containing deployment.yaml and service.yaml (required) - --cluster, -c: GKE cluster name (required) - --project, -p: GCP project ID (required) - --zone, -z: GKE cluster zone - --region, -r: GKE cluster region

Example:

deployml gke-deploy \
  -d ./manifests/mlflow \
  -c my-cluster \
  -p my-project-id \
  -z us-west1-a

`deployml gke-apply`

Apply previously generated manifests to GKE cluster.

Usage:

deployml gke-apply --config-path gke-config.yaml

Options: - --config-path, -c: Path to YAML config file (required) - --yes, -y: Skip confirmation prompts

Example:

# Apply manifests after editing
deployml gke-apply -c gke-config.yaml

Minikube Commands

`deployml minikube-init`

Initialize minikube and generate FastAPI Kubernetes manifests.

Usage:

deployml minikube-init \
  --output-dir ./manifests/fastapi \
  --image fastapi-demo:latest \
  --mlflow-uri http://mlflow-service:5000

Options: - --output-dir, -o: Directory to create Kubernetes manifests (required) - --image, -i: FastAPI Docker image (required) - --mlflow-uri, -m: MLflow tracking URI (optional) - --start-cluster: Start minikube cluster if not running (default: true)

Example:

deployml minikube-init \
  -o ./manifests/fastapi \
  -i fastapi-demo:latest \
  -m http://mlflow-service:5000

`deployml minikube-deploy`

Deploy FastAPI to minikube using kubectl apply.

Usage:

deployml minikube-deploy --manifest-dir ./manifests/fastapi

Options: - --manifest-dir, -d: Directory containing deployment.yaml and service.yaml (required) - --image-name, -i: Docker image name to load into minikube (auto-detected if not provided)

Example:

deployml minikube-deploy -d ./manifests/fastapi

`deployml mlflow-init`

Initialize minikube and generate MLflow Kubernetes manifests.

Usage:

deployml mlflow-init \
  --output-dir ./manifests/mlflow \
  --image mlflow-demo:latest \
  --backend-store-uri sqlite:///mlflow.db

Options: - --output-dir, -o: Directory to create Kubernetes manifests (required) - --image, -i: MLflow Docker image (required) - --backend-store-uri, -b: Backend store URI (defaults to SQLite) - --artifact-root, -a: Artifact root path (defaults to /mlflow-artifacts) - --start-cluster: Start minikube cluster if not running (default: true)

Example:

deployml mlflow-init \
  -o ./manifests/mlflow \
  -i mlflow-demo:latest \
  -b sqlite:///mlflow.db

`deployml mlflow-deploy`

Deploy MLflow to minikube using kubectl apply.

Usage:

deployml mlflow-deploy --manifest-dir ./manifests/mlflow

Options: - --manifest-dir, -d: Directory containing deployment.yaml and service.yaml (required) - --image-name, -i: Docker image name to load into minikube (auto-detected if not provided)

Example:

deployml mlflow-deploy -d ./manifests/mlflow

Teardown Commands

`deployml teardown`

Manage auto-teardown: cancel, status, update, or schedule.

Usage:

# Cancel scheduled teardown
deployml teardown cancel --config-path config.yaml

# Check teardown status
deployml teardown status --config-path config.yaml

# Update teardown schedule
deployml teardown update --config-path config.yaml

# Schedule new teardown
deployml teardown schedule --config-path config.yaml

Options: - --config-path, -c: Path to YAML config file (required)

Example:

# Check teardown status
deployml teardown status -c config.yaml

# Cancel teardown
deployml teardown cancel -c config.yaml

Complete Examples

Complete GKE Deployment

# 1. Generate manifests
deployml deploy --config-path gke-config.yaml --generate-only

# 2. Review manifests
ls -la .deployml/gke-mlflow-fastapi/manifests/

# 3. Apply manifests
deployml gke-apply --config-path gke-config.yaml

Cloud Run with Auto-Teardown

# Deploy with 24-hour auto-teardown
deployml deploy --config-path cloud-run-config.yaml

# Check teardown status
deployml teardown status --config-path cloud-run-config.yaml

# Cancel teardown
deployml teardown cancel --config-path cloud-run-config.yaml

Minikube Local Development

# 1. Start minikube
minikube start

# 2. Generate MLflow manifests
deployml mlflow-init -o ./manifests/mlflow -i mlflow-demo:latest

# 3. Deploy MLflow
deployml mlflow-deploy -d ./manifests/mlflow

# 4. Generate FastAPI manifests
deployml minikube-init -o ./manifests/fastapi -i fastapi-demo:latest -m http://mlflow-service:5000

# 5. Deploy FastAPI
deployml minikube-deploy -d ./manifests/fastapi

Next Steps

Read Installation Guide
Explore Tutorials