Skip to content

FLY_DEPLOYMENT.md

Latest commit

 

History

History
210 lines (157 loc) · 3.27 KB

FLY_DEPLOYMENT.md

File metadata and controls

210 lines (157 loc) · 3.27 KB

Deploying Cascade to Fly.io

Simple, cost-effective deployment using SQLite and local task execution.

Prerequisites

  1. Install Fly CLI:

    curl -L https://fly.io/install.sh | sh

  2. Sign up and log in:

    fly auth signup  # Or: fly auth login

Initial Deployment

  1. Create the app:

    fly apps create cascade  # Or use your preferred name

  2. Create persistent volume for SQLite:

    fly volumes create cascade_data --region sea --size 1

  3. Set secrets:

    # Generate and set secret key
fly secrets set SECRET_KEY_BASE=$(openssl rand -base64 64)

  4. Deploy:

    fly deploy

  5. Open the app:

    fly open

Configuration

Regions

Change primary_region in fly.toml to your preferred region:

  • sea - Seattle (US West)
  • iad - Ashburn (US East)
  • lhr - London
  • fra - Frankfurt
  • syd - Sydney

Full list

Scaling

Vertical scaling (change VM size):

# Increase memory
fly scale memory 1024  # 1GB RAM

# Increase CPUs
fly scale vm shared-cpu-2x  # 2x shared CPU

Horizontal scaling (add machines):

# Update fly.toml:
[auto_scaling]
  min_count = 1  # Keep 1 always running
  max_count = 4  # Scale up to 4 machines

Cost Optimization

Free tier (good for testing):

[vm]
  memory_mb = 256

[auto_scaling]
  min_count = 0  # Scale to zero when idle

Production (always-on):

[vm]
  cpu_kind = "shared"
  memory_mb = 1024

[auto_scaling]
  min_count = 1  # Keep 1 running
  max_count = 2  # Add redundancy

Adding DAGs

Upload DAG files to your deployed app:

# SSH into the running machine
fly ssh console

# Copy DAG files
fly ssh sftp shell
put my_dag.json /app/dags/

Or mount a local directory during development:

fly volumes create cascade_dags --size 1

Update fly.toml:

[[mounts]]
  source = "cascade_dags"
  destination = "/app/dags"

Monitoring

# View logs
fly logs

# Check status
fly status

# SSH into machine
fly ssh console

# Monitor metrics
fly dashboard

Database Access

# SSH into machine
fly ssh console

# Access database
cd /data
ls -lh cascade.db

# Copy database locally
fly ssh sftp get /data/cascade.db ./cascade_backup.db

Costs

Typical setup:

  • 1 shared-cpu-1x (512MB): ~$5/month
  • 1GB persistent volume: ~$0.15/month
  • Total: ~$5-6/month

Free tier eligible:

  • Scale to zero when idle: $0/month
  • Limited to 256MB RAM, slower startup

Updating

# Deploy new version
fly deploy

# Rollback if needed
fly releases
fly releases rollback <version>

Environment Variables

Set additional env vars:

fly secrets set CASCADE_WORKERS=8
fly secrets set DAGS_SCAN_INTERVAL=60

Troubleshooting

Container crashes:

fly logs --verbose

Database locked:

# SQLite doesn't support multiple writers
# Ensure only 1 machine is writing:
fly scale count 1

Out of memory:

fly scale memory 1024

Migration from Docker Compose

Your existing DAG files work as-is! Just:

  1. Deploy to Fly.io
  2. Copy DAG files to /app/dags
  3. Access at your Fly.io URL

No code changes needed!