Every company needs a deployment process, but every company is a bit different. As a company where developers own the Kubernetes config of their application, the best practice of separating the config from the code introduces unnecessary complexity. For instance, adding an environment variable requires coordination across multiple pull requests in multiple repositories with different deployment pipelines.

This talk will present the product process that led us to discover our development needs and the architecture work to create a deployment process based on ArgoCD and one repository for the code and the config.