Fly.io deployment configuration and scaling

# fly.toml
app = 'my-fastapi-app'
primary_region = 'ord'  # Chicago — pick closest to users

[build]
  dockerfile = 'Dockerfile'

[env]
  APP_ENV = 'production'
  PORT = '8000'
  # Non-secret config here

[http_service]
  internal_port = 8000
  force_https = true
  auto_stop_machines = true    # Stop when no traffic
  auto_start_machines = true   # Start on request
  min_machines_running = 0     # Can scale to zero (saves money)
  processes = ['app']

  [http_service.concurrency]
    type = 'requests'
    hard_limit = 100   # Max concurrent requests per machine
    soft_limit = 80    # Start new machine when this is hit

[[http_service.checks]]
  grace_period = '10s'
  interval = '15s'
  method = 'GET'
  path = '/health'
  timeout = '5s'

[mounts]
  # Persistent storage for file uploads
  source = 'uploads'
  destination = '/app/uploads'

[[vm]]
  size = 'shared-cpu-1x'  # 256MB RAM — good for APIs
  memory = '512mb'
  cpu_kind = 'shared'
  cpus = 1

# Deploy
fly deploy

# Set secrets (encrypted at rest, injected as env vars)
fly secrets set \
  DATABASE_URL="postgresql+asyncpg://user:pass@host/db" \
  OPENAI_API_KEY="sk-..." \
  REDIS_URL="redis://..."

# Scale machines manually
fly scale count 2 --region ord
fly scale vm performance-2x

# View logs
fly logs --app my-fastapi-app

# Open postgres console
fly postgres connect -a my-postgres-app

# Create persistent volume
fly volumes create uploads --region ord --size 10  # 10GB