Blossom Logo Blossom

Server Roles

Categories: configuration deployment infrastructure procfile
2 min read

Server Roles Reference

Server roles determine which processes run on each server. Roles map directly to process types defined in your application’s Procfile.

How Roles Work

If your Procfile contains:

web: bundle exec puma -C config/puma.rb --port $PORT
worker: bundle exec rake solid_queue:start
clock: bundle exec clockwork config/clock.rb

Then:

  • A server with the web role runs the Puma web server
  • A server with the worker role runs background job processing
  • A server with the clock role runs scheduled tasks
  • You can assign multiple roles to one server (e.g., web,worker)

Role Types

Infrastructure Roles - Handle infrastructure tasks, no application code deployed:

  • load balancer: Routes traffic to application servers
  • build: Dedicated server for building Docker images
  • db/database: Database server
  • service: Generic infrastructure service

Application Roles - Run your application processes:

  • web: Web application processes
  • worker: Background jobs
  • clock: Scheduled tasks
  • Any custom role matching your Procfile processes

Key Rules

  • Servers with only infrastructure roles won’t receive application code
  • Servers need at least one application role to receive deployments
  • You can combine multiple roles on a single server
  • Role names must match process names in your Procfile

Recommendations

Database Servers: Use dedicated servers for databases. Databases are stateful - losing a database server means losing data. Don’t combine database roles with other roles.

Application Servers: These are stateless and can be safely replaced. Feel free to combine roles for cost optimization.

Build Servers: Use dedicated build servers to prevent Docker builds from impacting application performance.

Load Balancers: Consider dedicated load balancers for production environments to ensure reliable traffic distribution.

Server Architecture Services
Blossom Docs