What problem does it solve? Writing Grafana Alloy pipeline configuration by hand is error-prone: undocumented attributes slip through, ref-valued fields get quoted as strings, and alloy validate misses type mismatches that only fail at load time. This Skill provides the conventions, schema map, and invariants for Materialize's pipelines-as-code system, where YAML pipeline definitions are validated against embedded JSONSchemas and rendered to canonical .alloy config via typed Rust structs. ## Core Features & Use Cases - Typed pipeline authoring: Write log-processing (loki.process stages, discovery.kubernetes, relabel rules) and metric-processing (prometheus.*, otelcol.*) pipelines in YAML under packages/alloy-pipelines/, validated by schemas in packages/mzmon-lib/schemas/alloy/. - Schema extension guidance: Add new components, stages, attributes, or typed sub-blocks by extending the JSONSchemas and the Rust ComponentBlock sugar in lockstep, with the raw: escape hatch as a deliberate fallback. - Debugging rendered output: Diagnose mz-monitoring-build gen-pipelines failures, serde/validation drift, and capsule type mismatches using documented load-bearing invariants. - Use Case: When adding a new otelcol.processor to the gateway metrics pipeline, use this Skill to extend otelcol.schema.yaml, add the typed struct in components/otelcol.rs, wire it into gateway.yaml, and verify with make pipelines and cargo test -p mzmon-lib. ## Quick Start Ask the AI to add a new typed attribute or component to the Alloy pipeline schemas and render the updated gateway pipeline with make pipelines.
