Your Helm Zoo Will Kill You

Helm, a popular tool for managing Kubernetes applications, has fundamental design flaws that can lead to dramatic long-term consequences, particularly in enterprise environments with large teams and complex applications. The current approach to writing charts and managing resources leads to a "chart zoo" problem, where it's difficult to track changes, maintain, and roll-back updates. This is due to the lack of standardization and proper management of out-of-cluster resources. A new approach is proposed, which involves treating manifests as a protocol, standardizing changes through an API, and managing out-of-cluster resources through deployment sets. By following this design pattern, teams can achieve clear tracking of changes, standardized charts, reduced overhead, and improved developer productivity.