Skip to main content

Key conceptual difference

Render’s approach: all-in-one service definitions

In Render, each service is a complete, self-contained deployment unit. If you want the same application in test and production, you create two separate service definitions with duplicated build commands, even though the code and build process are identical. This leads to repetition and makes it harder to maintain consistency across environments.

Ryvn’s approach: separation of concerns

Ryvn separates what you’re building (services) from how you’re deploying it (installations):
  • Services: Define the source code, repository, and build process once
  • Installations: Configure deployment-specific settings like resources, scaling, and environment variables per environment
Why this matters: Guaranteed consistency - The exact same container image that passes tests in staging is what runs in production. No more “works on my machine” or subtle build differences between environments. No more copy-paste - When you need to update your build process, change it in one place. No hunting through dozens of duplicated service definitions to keep them in sync.

Quick mapping reference

Service definition (build time)

Installation configuration (deploy time)

Environment variables

Web service example

Ingress networking and NGINX annotations are deprecated. See Ingress networking for what to use instead.

Cron job example

Migration steps

  1. Create variable groups for your Render environment groups
  2. Define services as resource YAML files with build configuration
  3. Add environments with provider configuration
  4. Create installations per environment with runtime config
  5. Configure Git Sync in Ryvn Dashboard
  6. Push and deploy
Key differences to remember:
  • Repository format: owner/repo not full GitHub URL
  • Build commands go in service definition
  • Start commands go in installation config
  • Resources must be explicit (no “plan” names)
  • Web services need explicit ingress configuration