fix(otlp): composed schema shared by all build consumers

Author: lucastemb

lib/datadog-agent/config-overlay-model/src/schema_gen.rs

@@ -94,6 +94,14 @@ pub struct FieldInfo {
 pub fn load_schema(datadog_schema: &Path, otel_schema_dir: &Path) -> IndexMap<String, FieldInfo> {
     let doc = crate::load_composed_schema(datadog_schema, otel_schema_dir)
         .unwrap_or_else(|e| panic!("failed to load composed schema: {e}"));
+    load_schema_from_value(&doc)
+}
+
+/// Builds the flat `yaml_path → FieldInfo` map from a pre-loaded composed schema.
+///
+/// Callers that already hold the composed schema should use this instead of [`load_schema`] to
+/// avoid a redundant disk read and guarantee every consumer sees the same document.
+pub fn load_schema_from_value(doc: &serde_yaml::Value) -> IndexMap<String, FieldInfo> {
     let properties = doc
         .get("properties")
         .and_then(|v| v.as_mapping())

lib/datadog-agent/config/build.rs

@@ -1,6 +1,6 @@
 use std::path::PathBuf;
 
-use datadog_agent_config_overlay_model::{schema_gen, Files, SchemaOverlay};
+use datadog_agent_config_overlay_model::{load_composed_schema, schema_gen, Files, SchemaOverlay};
 
 #[path = "build/classifier_gen.rs"]
 mod classifier_gen;
@@ -28,12 +28,14 @@ fn main() {
     println!("cargo:rerun-if-changed=build/env_reader_gen.rs");
     println!("cargo:rerun-if-changed=build/witness_gen.rs");
 
-    let schema_path = files.datadog_schema.clone();
-    let schema_map = schema_gen::load_schema(&files.datadog_schema, &files.otel_schema_dir);
+    // Load the composed schema once and pass the in-memory value to every consumer.
+    let composed_schema = load_composed_schema(&files.datadog_schema, &files.otel_schema_dir)
+        .unwrap_or_else(|e| panic!("failed to load composed schema: {e}"));
+    let schema_map = schema_gen::load_schema_from_value(&composed_schema);
     let overlay = SchemaOverlay::load(files).unwrap_or_else(|e| panic!("{e}"));
 
     classifier_gen::generate(&overlay, &schema_map, &manifest_dir);
-    datadog_config_gen::generate(&overlay, &schema_path, &schema_map, &manifest_dir);
+    datadog_config_gen::generate(&overlay, &composed_schema, &schema_map, &manifest_dir);
     env_reader_gen::generate(&overlay, &schema_map, &manifest_dir);
     // Must run after datadog_config_gen: it parses the freshly-written datadog_configuration.rs.
     witness_gen::generate(&overlay, &manifest_dir);

lib/datadog-agent/config/build/datadog_config_gen.rs

@@ -24,7 +24,7 @@ use std::collections::{BTreeMap, HashMap, HashSet};
 use std::path::Path;
 
 use datadog_agent_config_overlay_model::schema_gen::{FieldInfo, FieldType};
-use datadog_agent_config_overlay_model::{load_resolved_schema, InputShape, KnownEntry, SchemaOverlay};
+use datadog_agent_config_overlay_model::{InputShape, KnownEntry, SchemaOverlay};
 use indexmap::IndexMap;
 use serde_json::{Map, Value};
 use syn::visit_mut::{self, VisitMut};
@@ -43,16 +43,13 @@ const LEAF_KEYWORDS: &[&str] = &[
 ];
 
 pub fn generate(
-    overlay: &SchemaOverlay, schema_path: &Path, schema_map: &IndexMap<String, FieldInfo>, manifest_dir: &Path,
+    overlay: &SchemaOverlay, composed_schema: &serde_yaml::Value, schema_map: &IndexMap<String, FieldInfo>,
+    manifest_dir: &Path,
 ) {
     let supported = supported_keys(overlay);
 
-    // Load the schema with all `$ref` files inlined (apm_config, multi_region_failover, ...).
-    // Without $ref resolution those subsystem keys are silently absent and the witness driver
-    // cannot cover them. Delegate to overlay-model's resolver rather than duplicating it.
-    let schema_yaml = load_resolved_schema(schema_path).unwrap_or_else(|e| panic!("failed to load schema: {e}"));
     let schema: Value =
-        serde_json::to_value(schema_yaml).unwrap_or_else(|e| panic!("failed to convert schema to JSON: {e}"));
+        serde_json::to_value(composed_schema).unwrap_or_else(|e| panic!("failed to convert schema to JSON: {e}"));
     let properties = schema
         .get("properties")
         .and_then(Value::as_object)