Skip to content

DEPLOYMENT

🚀 Deployment Patterns

Common deployment patterns and best practices for Aether workloads.

📑 Table of Contents

📦 Single Workload Deployment

The simplest pattern: one workload spec, one deployment.

# Validate → Build → Deploy
aether validate
aether build
aether run

# Or with runtime override
aether run --runtime kubernetes

# Dry-run to preview
aether --dry-run run --runtime kube

Lifecycle

aether status my-app     # Check status
aether logs my-app       # View logs
aether logs my-app -f    # Follow logs
aether exec my-app       # Shell into container
aether stop my-app       # Stop
aether delete my-app     # Remove completely

📁 Directory Batch Deploy

Deploy all workload specs from a directory at once.

# Deploy all YAML files in specs/
aether deploy ./specs/

# With runtime override
aether deploy ./specs/ --runtime kubernetes

# Stop on first failure
aether deploy ./specs/ --fail-fast

# Preview without deploying
aether deploy ./specs/ --dry-run

The deploy command: 1. Discovers all *.yaml files in the directory 2. Validates each spec 3. Deploys in parallel (or sequentially with --fail-fast) 4. Reports results summary

🎼 Compose Multi-Workload Stacks

For applications with multiple services and dependencies.

Compose File

# aether-compose.yaml
version: "1"
workloads:
  frontend:
    spec: ./frontend/workload.yaml
    runtime: podman
    depends_on:
      - backend
    environment:
      API_URL: http://backend:8080
  backend:
    spec: ./backend/workload.yaml
    runtime: kubernetes
    depends_on:
      - database
  database:
    spec: ./database/workload.yaml
    runtime: kubernetes

Commands

aether compose validate              # Validate compose file
aether compose up                    # Deploy all (respects dependency order)
aether compose up --runtime kube     # Override runtime for all
aether compose up --dry-run          # Preview
aether compose down                  # Stop all

📝 Template-Based Deployment

Generate workload specs from 8 built-in templates.

# List available templates
aether template --list

# Generate from template
aether template web-app --workload-name my-site --owner team --project demo

# Save to file
aether template rest-api --workload-name api-svc -o api.yaml

# Then deploy
aether -s api.yaml run

Available Templates

Template Ports Use Case
web-app 80, 443 Web frontends
rest-api 8080 API backends
database 5432 PostgreSQL, MySQL
cache 6379 Redis, Memcached
worker — Background jobs
cron-job — Scheduled tasks
ml-training — GPU workloads
microservice 8080 Generic microservice

🔄 Environment Promotion

Promote workloads through environments (dev → staging → production).

# Create environments
aether env create dev --tier development
aether env create staging --tier staging
aether env create prod --tier production

# Promote
aether env promote my-app dev staging
aether env promote my-app staging prod

# Check parity
aether env parity staging prod

🔁 CI/CD Pipeline Deployment

GitHub Actions

- name: Deploy
  run: |
    aether validate -s workload.yaml
    aether policy-check -s workload.yaml --policy production
    aether --yes run -s workload.yaml --runtime kubernetes

Key Flags for CI

Flag Purpose
--yes Auto-confirm prompts
--json Machine-readable output
--quiet Suppress non-error output
--dry-run Preview without executing
--output json JSON output format

👁 Watch Mode (Hot Reload)

Auto-redeploy when spec file changes (development).

aether watch
aether watch --runtime podman

Watch mode: 1. Polls the spec file every 500ms 2. Debounces changes (waits 1s after last change) 3. Reads file atomically to prevent partial reads 4. Validates, then redeploys

Document Description
Quick Start 5-minute getting started
Compose Guide Multi-workload deployment details
CI/CD Guide Pipeline integration
Templates Template catalog