Expand description
Configuration key classifier.
A programmatic registry of all recognized configuration keys. Each entry describes the key purely from the configuration system’s perspective: its canonical YAML path, the environment variables that map to it, the shape of its value, and which internal config structs consume it.
This registry is intentionally free of Rust field names and struct internals—it models the configuration surface as an operator would see it, and can be used at runtime to detect unknown or unsupported keys in a loaded configuration file.
§User Guide
The registry is generated at build time from schema_overlay.yaml, which partitions every
key in core_schema.yaml into exactly one of: supported, unsupported, or ignored. Data
integrity (uniqueness, full coverage, sorted sections) is enforced by SchemaOverlay::load()
during the build.
§Adding a Configuration Key
Add a supported entry in lib/datadog-agent/config/schema/schema_overlay.yaml with
support_level, pipelines, used_by, description, and config_registry_filename.
The build generates the annotation constants automatically.
§Updating the Vendored Schema
After updating core_schema.yaml, the build will fail if any new keys are not covered
by the overlay. For each new key, add it to the appropriate section of
schema_overlay.yaml.
Modules§
- structs
- Identifiers for known configuration structs.
Structs§
- Classification
- Result of classifying a single config key/value pair against the registry.
- Classifier
Entry - Slim per-key data generated at build time for the classifier.
- Config
Classifier - Classifies the support level of config keys and determines if a value is default.
Enums§
- Pipeline
- The ADP pipeline a config key affects.
- Pipeline
Affinity - Which pipelines a config key affects.
- Severity
- The
Severitylevel of a config key that Saluki doesn’t support. - Support
Level - The support level for a given configuration key.