chore(llmobs): update evaluation metric api endpoint to v2 (#7511)

sabrenner · web-flow · commit c589ad403bbf · 2026-02-12T15:07:02.000-05:00
* use v2 api endpoint

* update tests
diff --git a/packages/dd-trace/src/llmobs/constants/writers.js b/packages/dd-trace/src/llmobs/constants/writers.js
@@ -10,7 +10,7 @@ module.exports = {
 
   EVALUATIONS_INTAKE: 'api',
   EVALUATIONS_EVENT_TYPE: 'evaluation_metric',
-  EVALUATIONS_ENDPOINT: '/api/intake/llm-obs/v1/eval-metric',
+  EVALUATIONS_ENDPOINT: '/api/intake/llm-obs/v2/eval-metric',
 
   EVP_PAYLOAD_SIZE_LIMIT: 5 << 20, // 5MB (actual limit is 5.1MB)
   EVP_EVENT_SIZE_LIMIT: (1 << 20) - 1024, // 999KB (actual limit is 1MB)
diff --git a/packages/dd-trace/src/llmobs/sdk.js b/packages/dd-trace/src/llmobs/sdk.js
@@ -412,8 +412,12 @@ class LLMObs extends NoopLLMObs {
       }
 
       const payload = {
-        span_id: spanId,
-        trace_id: traceId,
+        join_on: {
+          span: {
+            span_id: spanId,
+            trace_id: traceId,
+          },
+        },
         label,
         metric_type: metricType,
         ml_app: mlApp,
diff --git a/packages/dd-trace/test/llmobs/sdk/index.spec.js b/packages/dd-trace/test/llmobs/sdk/index.spec.js
@@ -1189,8 +1189,12 @@ describe('sdk', () => {
       })
 
       assert.deepStrictEqual(LLMObsEvalMetricsWriter.prototype.append.getCall(0).args[0], {
-        trace_id: spanCtx.traceId,
-        span_id: spanCtx.spanId,
+        join_on: {
+          span: {
+            trace_id: spanCtx.traceId,
+            span_id: spanCtx.spanId,
+          },
+        },
         ml_app: 'test',
         timestamp_ms: 1234,
         label: 'test',
@@ -1241,8 +1245,12 @@ describe('sdk', () => {
       const evalMetric = LLMObsEvalMetricsWriter.prototype.append.getCall(0).args[0]
 
       assert.deepStrictEqual(evalMetric, {
-        span_id: '5678',
-        trace_id: '1234',
+        join_on: {
+          span: {
+            span_id: '5678',
+            trace_id: '1234',
+          },
+        },
         label: 'has_toxicity',
         metric_type: 'boolean',
         ml_app: 'mlApp',
diff --git a/packages/dd-trace/test/llmobs/sdk/integration.spec.js b/packages/dd-trace/test/llmobs/sdk/integration.spec.js
@@ -142,8 +142,12 @@ describe('end to end sdk integration tests', () => {
       assert.equal(llmobsEvaluationMetrics.length, 1)
 
       assertLlmObsEvaluationMetric(llmobsEvaluationMetrics[0], {
-        traceId: llmobsSpans[0].trace_id,
-        spanId: llmobsSpans[0].span_id,
+        joinOn: {
+          span: {
+            traceId: llmobsSpans[0].trace_id,
+            spanId: llmobsSpans[0].span_id,
+          },
+        },
         label: 'foo',
         metricType: 'categorical',
         mlApp: 'test',
diff --git a/packages/dd-trace/test/llmobs/util.js b/packages/dd-trace/test/llmobs/util.js
@@ -40,8 +40,12 @@ const MOCK_NOT_NULLISH = Symbol('not-nullish')
 /**
  * @typedef {{
  *   label: string,
- *   traceId: string,
- *   spanId: string,
+ *   joinOn: {
+ *     span: {
+ *       traceId: string,
+ *       spanId: string,
+ *     },
+ *   },
  *   metricType: 'categorical' | 'score',
  *   mlApp: string,
  *   timestamp?: number,
@@ -279,8 +283,12 @@ function assertLlmObsSpanEvent (actual, expected) {
 function assertLlmObsEvaluationMetric (actual, expected) {
   const {
     label,
-    traceId = MOCK_STRING,
-    spanId = MOCK_STRING,
+    joinOn = {
+      span: {
+        traceId: MOCK_STRING,
+        spanId: MOCK_STRING,
+      },
+    },
     metricType,
     mlApp,
     timestamp = MOCK_NUMBER,
@@ -318,8 +326,12 @@ function assertLlmObsEvaluationMetric (actual, expected) {
   }
 
   const expectedEvaluationMetric = {
-    span_id: spanId,
-    trace_id: traceId,
+    join_on: {
+      span: {
+        trace_id: joinOn.span.traceId,
+        span_id: joinOn.span.spanId,
+      },
+    },
     label,
     metric_type: metricType,
     ml_app: mlApp,
diff --git a/packages/dd-trace/test/llmobs/writers/evaluations.spec.js b/packages/dd-trace/test/llmobs/writers/evaluations.spec.js
@@ -29,7 +29,7 @@ describe('LLMObsEvalMetricsWriter', () => {
 
     writer.flush = flush // just to stop the beforeExit flush call
 
-    assert.strictEqual(writer.url, 'https://api.datadoghq.com/api/intake/llm-obs/v1/eval-metric')
+    assert.strictEqual(writer.url, 'https://api.datadoghq.com/api/intake/llm-obs/v2/eval-metric')
     assert.strictEqual(writer._eventType, 'evaluation_metric')
   })
 
@@ -39,7 +39,7 @@ describe('LLMObsEvalMetricsWriter', () => {
       hostname: 'localhost',
     })
     writer.setAgentless(false)
-    assert.strictEqual(writer.url, 'http://localhost:8126/evp_proxy/v2/api/intake/llm-obs/v1/eval-metric')
+    assert.strictEqual(writer.url, 'http://localhost:8126/evp_proxy/v2/api/intake/llm-obs/v2/eval-metric')
     assert.strictEqual(writer._eventType, 'evaluation_metric')
   })
 
diff --git a/packages/dd-trace/test/plugins/agent.js b/packages/dd-trace/test/plugins/agent.js
@@ -473,7 +473,7 @@ module.exports = {
     })
 
     // LLM Observability evaluation metrics endpoint
-    agent.post('/evp_proxy/v2/api/intake/llm-obs/v1/eval-metric', (req, res) => {
+    agent.post('/evp_proxy/v2/api/intake/llm-obs/v2/eval-metric', (req, res) => {
       llmobsEvaluationMetricsRequests.push(JSON.parse(req.body))
       res.status(200).send()
     })
