KYAML: A Stricter YAML Dialect for Kubernetes Manifests

The Challenge with Standard YAML in Kubernetes

YAML has been the standard for Kubernetes manifests, but its flexibility leads to inconsistencies and potential issues. While not inherently a bad format, YAML offers many choices, some of which make Kubernetes configuration files harder to read or prone to surprising behavior due to features Kubernetes does not fully utilize.

Introducing KYAML for Consistency

Recognizing that Kubernetes only relies on a small subset of YAML's features, SIG CLI introduced KYAML. This is not a new configuration language but a stricter, more consistent way to write YAML, effectively an agreed-upon style. It standardizes on the necessary parts of YAML, avoiding the rest, and is designed to be parseable by the existing ecosystem without changes, as proposed in KEP 5295.

Addressing Common YAML Pitfalls

Standard YAML has well-known traps, such as whitespace sensitivity, where incorrect indentation can alter the object's representation. Another issue is silent type coercion, where values that appear to be strings are implicitly converted to other types, exemplified by the 'Norway Bug' where 'NO' is parsed as a boolean 'false'. KYAML addresses these by making structure and types explicit.

How KYAML Improves Manifests

KYAML explicitly defines structure without relying on whitespace, always quotes value strings to prevent silent type coercion, and consistently uses '{}' for maps and structs and '[]' for lists. It also supports comments, offering a more robust and predictable experience for writing Kubernetes configuration files compared to standard YAML or JSON, which lacks comment support and has strict formatting requirements.