feat(sampling): accept Remote Config list-shape tags natively

bm1549 · claude · bm1549 · commit cdac292e6702 · 2026-05-22T10:43:51.000-04:00
SamplingRuleConfig::tags now accepts both the map shape
({"env": "prod"}) and the Remote Config wire shape
([{"key": "env", "value_glob": "prod"}]). List entries missing
"key" or "value_glob" produce a deserialization error — we don't
silently drop entries because doing so could broaden a tag-
constrained rule.

This removes the per-tracer normalize_rc_tags workaround. The
dd-trace-rs caller stops normalizing in a follow-up PR once
this lands in a libdatadog release.

Co-Authored-By: Claude Opus 4.7 &lt;noreply@anthropic.com&gt;
diff --git a/libdd-sampling/src/sampling_rule_config.rs b/libdd-sampling/src/sampling_rule_config.rs
@@ -25,8 +25,13 @@ pub struct SamplingRuleConfig {
     #[serde(default)]
     pub resource: Option<String>,
 
-    /// Tags that must match (key-value pairs)
-    #[serde(default)]
+    /// Tags that must match (key-value pairs).
+    ///
+    /// Accepts either the map shape `{"env": "prod"}` or the Remote Config
+    /// wire shape `[{"key": "env", "value_glob": "prod"}]`. Internally both
+    /// normalize to the map shape; the list-shape entries are required to
+    /// have both `key` and `value_glob` (missing either rejects the rule).
+    #[serde(default, deserialize_with = "deserialize_tags")]
     pub tags: HashMap<String, String>,
 
     /// Where this rule comes from (customer, dynamic, default).
@@ -61,6 +66,61 @@ fn default_provenance() -> String {
     "default".to_string()
 }
 
+/// Deserializes the `tags` field, accepting either:
+///   - map shape:  `{"env": "prod", "region": "us-east-1"}`
+///   - list shape: `[{"key": "env", "value_glob": "prod"}, ...]`
+///
+/// A list entry missing `key` or `value_glob` produces a deserialization
+/// error; we never silently drop entries because that could broaden a
+/// tag-constrained sampling rule.
+fn deserialize_tags<'de, D>(deserializer: D) -> Result<HashMap<String, String>, D::Error>
+where
+    D: serde::Deserializer<'de>,
+{
+    use serde::de::{MapAccess, SeqAccess, Visitor};
+    use std::fmt;
+
+    #[derive(serde::Deserialize)]
+    struct ListEntry {
+        key: String,
+        value_glob: String,
+    }
+
+    struct TagsVisitor;
+
+    impl<'de> Visitor<'de> for TagsVisitor {
+        type Value = HashMap<String, String>;
+
+        fn expecting(&self, f: &mut fmt::Formatter) -> fmt::Result {
+            f.write_str("a map of string to string or a list of {key, value_glob} objects")
+        }
+
+        fn visit_map<M>(self, mut access: M) -> Result<Self::Value, M::Error>
+        where
+            M: MapAccess<'de>,
+        {
+            let mut map = HashMap::with_capacity(access.size_hint().unwrap_or(0));
+            while let Some((k, v)) = access.next_entry::<String, String>()? {
+                map.insert(k, v);
+            }
+            Ok(map)
+        }
+
+        fn visit_seq<S>(self, mut access: S) -> Result<Self::Value, S::Error>
+        where
+            S: SeqAccess<'de>,
+        {
+            let mut map = HashMap::with_capacity(access.size_hint().unwrap_or(0));
+            while let Some(entry) = access.next_element::<ListEntry>()? {
+                map.insert(entry.key, entry.value_glob);
+            }
+            Ok(map)
+        }
+    }
+
+    deserializer.deserialize_any(TagsVisitor)
+}
+
 #[derive(Debug, Default, Clone, PartialEq)]
 pub struct ParsedSamplingRules {
     pub rules: Vec<SamplingRuleConfig>,
@@ -181,6 +241,65 @@ mod tests {
         assert_eq!(original, restored);
     }
 
+    #[test]
+    fn test_sampling_rule_config_tags_accepts_map_shape() {
+        // Already supported — guard against regression.
+        let json = r#"{
+            "sample_rate": 0.5,
+            "service": "svc",
+            "tags": {"env": "prod", "region": "us-east-1"}
+        }"#;
+        let cfg: SamplingRuleConfig = serde_json::from_str(json).unwrap();
+        assert_eq!(cfg.tags.get("env").map(String::as_str), Some("prod"));
+        assert_eq!(
+            cfg.tags.get("region").map(String::as_str),
+            Some("us-east-1")
+        );
+    }
+
+    #[test]
+    fn test_sampling_rule_config_tags_accepts_rc_list_shape() {
+        // Remote Config wire shape: list of {key, value_glob} entries.
+        let json = r#"{
+            "sample_rate": 0.5,
+            "service": "svc",
+            "tags": [
+                {"key": "env", "value_glob": "prod"},
+                {"key": "region", "value_glob": "us-east-1"}
+            ]
+        }"#;
+        let cfg: SamplingRuleConfig = serde_json::from_str(json).unwrap();
+        assert_eq!(cfg.tags.get("env").map(String::as_str), Some("prod"));
+        assert_eq!(
+            cfg.tags.get("region").map(String::as_str),
+            Some("us-east-1")
+        );
+    }
+
+    #[test]
+    fn test_sampling_rule_config_tags_list_with_malformed_entry_rejects() {
+        // A list entry missing `value_glob` must reject the whole rule rather
+        // than silently dropping the entry — silently dropping a constraint could
+        // broaden a tag-constrained rule and produce a security-relevant change
+        // in sampling decisions.
+        let json = r#"{
+            "sample_rate": 0.5,
+            "tags": [
+                {"key": "env", "value_glob": "prod"},
+                {"key": "region"}
+            ]
+        }"#;
+        let result: Result<SamplingRuleConfig, _> = serde_json::from_str(json);
+        assert!(result.is_err(), "expected deserialization to fail");
+    }
+
+    #[test]
+    fn test_sampling_rule_config_tags_absent_defaults_to_empty() {
+        let json = r#"{"sample_rate": 0.5}"#;
+        let cfg: SamplingRuleConfig = serde_json::from_str(json).unwrap();
+        assert!(cfg.tags.is_empty());
+    }
+
     #[test]
     fn test_sampling_rule_config_display() {
         let config = SamplingRuleConfig {
