@node-llm/monitor-otel 0.1.0 → 0.1.1

package/LICENSE

@@ -0,0 +1,21 @@
+MIT License
+
+Copyright (c) 2026 NodeLLM contributors
+
+Permission is hereby granted, free of charge, to any person obtaining a copy
+of this software and associated documentation files (the "Software"), to deal
+in the Software without restriction, including without limitation the rights
+to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
+copies of the Software, and to permit persons to whom the Software is
+furnished to do so, subject to the following conditions:
+
+The above copyright notice and this permission notice shall be included in all
+copies or substantial portions of the Software.
+
+THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
+AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
+OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
+SOFTWARE.

package/dist/index.js

@@ -75,6 +75,8 @@ function extractAIAttributes(span) {
     // Usage - AI SDK attributes
     promptTokens: getNumber(attrs, "ai.usage.promptTokens"),
     completionTokens: getNumber(attrs, "ai.usage.completionTokens"),
+    cachedInputTokens: getNumber(attrs, "ai.usage.cachedInputTokens"),
+    reasoningTokens: getNumber(attrs, "ai.usage.reasoningTokens"),
     // Timing (streaming)
     msToFirstChunk: getNumber(attrs, "ai.response.msToFirstChunk"),
     msToFinish: getNumber(attrs, "ai.response.msToFinish"),
@@ -90,6 +92,9 @@ function extractAIAttributes(span) {
     genAiResponseModel: getString(attrs, "gen_ai.response.model"),
     genAiInputTokens: getNumber(attrs, "gen_ai.usage.input_tokens"),
     genAiOutputTokens: getNumber(attrs, "gen_ai.usage.output_tokens"),
+    genAiCacheReadInputTokens: getNumber(attrs, "gen_ai.usage.cache_read.input_tokens"),
+    genAiCacheCreationInputTokens: getNumber(attrs, "gen_ai.usage.cache_creation.input_tokens"),
+    genAiReasoningOutputTokens: getNumber(attrs, "gen_ai.usage.reasoning.output_tokens"),
     // Custom metadata
     metadata: Object.keys(metadata).length > 0 ? metadata : void 0
   };
@@ -229,6 +234,9 @@ var NodeLLMSpanProcessor = class {
     const duration = calculateDurationMs(span.startTime, span.endTime);
     const promptTokens = attrs.promptTokens ?? attrs.genAiInputTokens ?? 0;
     const completionTokens = attrs.completionTokens ?? attrs.genAiOutputTokens ?? 0;
+    const cachedTokens = attrs.cachedInputTokens ?? attrs.genAiCacheReadInputTokens ?? 0;
+    const cacheCreationTokens = attrs.genAiCacheCreationInputTokens ?? 0;
+    const reasoningTokens = attrs.reasoningTokens ?? attrs.genAiReasoningOutputTokens ?? 0;
     const cost = estimateCost(provider, model, promptTokens, completionTokens);
     const requestId = this.options.sessionIdExtractor ? generateRequestId(span) : generateRequestId(span);
     const sessionId = this.options.sessionIdExtractor ? this.options.sessionIdExtractor(span) : extractSessionId(span);
@@ -239,7 +247,10 @@ var NodeLLMSpanProcessor = class {
       usage: {
         promptTokens,
         completionTokens,
-        totalTokens: promptTokens + completionTokens
+        totalTokens: promptTokens + completionTokens,
+        ...cachedTokens > 0 && { cachedTokens },
+        ...cacheCreationTokens > 0 && { cacheCreationTokens },
+        ...reasoningTokens > 0 && { reasoningTokens }
       },
       // Streaming metrics
       ...attrs.msToFirstChunk !== void 0 && {

package/package.json

@@ -1,20 +1,14 @@
 {
   "name": "@node-llm/monitor-otel",
-  "version": "0.1.0",
+  "version": "0.1.1",
   "description": "OpenTelemetry integration for NodeLLM Monitor",
   "main": "dist/index.js",
   "type": "module",
   "types": "dist/index.d.ts",
-  "scripts": {
-    "build": "tsup src/index.ts --format esm --dts --clean",
-    "dev": "tsup src/index.ts --format esm --dts --watch",
-    "lint": "eslint .",
-    "test": "vitest run"
-  },
   "dependencies": {
-    "@node-llm/monitor": "workspace:*",
     "@opentelemetry/api": "^1.7.0",
-    "@opentelemetry/sdk-trace-base": "^1.18.1"
+    "@opentelemetry/sdk-trace-base": "^1.18.1",
+    "@node-llm/monitor": "0.4.2"
   },
   "devDependencies": {
     "@types/node": "^25.2.0",
@@ -22,5 +16,11 @@
     "tsup": "^8.0.1",
     "typescript": "^5.3.3",
     "vitest": "^4.0.18"
+  },
+  "scripts": {
+    "build": "tsup src/index.ts --format esm --dts --clean",
+    "dev": "tsup src/index.ts --format esm --dts --watch",
+    "lint": "eslint .",
+    "test": "vitest run"
   }
-}
+}

package/src/NodeLLMSpanProcessor.ts

@@ -137,6 +137,9 @@ export class NodeLLMSpanProcessor implements OTelSpanProcessor {
     // Calculate token counts (prefer AI SDK attributes, fallback to GenAI conventions)
     const promptTokens = attrs.promptTokens ?? attrs.genAiInputTokens ?? 0;
     const completionTokens = attrs.completionTokens ?? attrs.genAiOutputTokens ?? 0;
+    const cachedTokens = attrs.cachedInputTokens ?? attrs.genAiCacheReadInputTokens ?? 0;
+    const cacheCreationTokens = attrs.genAiCacheCreationInputTokens ?? 0;
+    const reasoningTokens = attrs.reasoningTokens ?? attrs.genAiReasoningOutputTokens ?? 0;
 
     // Estimate cost
     const cost = estimateCost(provider, model, promptTokens, completionTokens);
@@ -159,7 +162,10 @@ export class NodeLLMSpanProcessor implements OTelSpanProcessor {
       usage: {
         promptTokens,
         completionTokens,
-        totalTokens: promptTokens + completionTokens
+        totalTokens: promptTokens + completionTokens,
+        ...(cachedTokens > 0 && { cachedTokens }),
+        ...(cacheCreationTokens > 0 && { cacheCreationTokens }),
+        ...(reasoningTokens > 0 && { reasoningTokens })
       },
       // Streaming metrics
       ...(attrs.msToFirstChunk !== undefined && {

package/src/types.ts

@@ -81,6 +81,9 @@ export interface AISpanAttributes {
   promptTokens?: number;
   completionTokens?: number;
   totalTokens?: number;
+  // Cache/reasoning usage (forward-compatible AI SDK attributes - not yet
+  cachedInputTokens?: number;
+  reasoningTokens?: number;
 
   // Timing (streaming)
   msToFirstChunk?: number;
@@ -99,6 +102,9 @@ export interface AISpanAttributes {
   genAiResponseModel?: string;
   genAiInputTokens?: number;
   genAiOutputTokens?: number;
+  genAiCacheReadInputTokens?: number;
+  genAiCacheCreationInputTokens?: number;
+  genAiReasoningOutputTokens?: number;
 
   // Custom metadata
   metadata?: Record<string, unknown>;

package/src/utils.ts

@@ -131,6 +131,8 @@ export function extractAIAttributes(span: OTelReadableSpan): AISpanAttributes {
     // Usage - AI SDK attributes
     promptTokens: getNumber(attrs, "ai.usage.promptTokens"),
     completionTokens: getNumber(attrs, "ai.usage.completionTokens"),
+    cachedInputTokens: getNumber(attrs, "ai.usage.cachedInputTokens"),
+    reasoningTokens: getNumber(attrs, "ai.usage.reasoningTokens"),
 
     // Timing (streaming)
     msToFirstChunk: getNumber(attrs, "ai.response.msToFirstChunk"),
@@ -149,6 +151,9 @@ export function extractAIAttributes(span: OTelReadableSpan): AISpanAttributes {
     genAiResponseModel: getString(attrs, "gen_ai.response.model"),
     genAiInputTokens: getNumber(attrs, "gen_ai.usage.input_tokens"),
     genAiOutputTokens: getNumber(attrs, "gen_ai.usage.output_tokens"),
+    genAiCacheReadInputTokens: getNumber(attrs, "gen_ai.usage.cache_read.input_tokens"),
+    genAiCacheCreationInputTokens: getNumber(attrs, "gen_ai.usage.cache_creation.input_tokens"),
+    genAiReasoningOutputTokens: getNumber(attrs, "gen_ai.usage.reasoning.output_tokens"),
 
     // Custom metadata
     metadata: Object.keys(metadata).length > 0 ? metadata : undefined

package/test/NodeLLMSpanProcessor.test.ts

@@ -72,6 +72,31 @@ describe("NodeLLMSpanProcessor", () => {
     expect(event.payload.usage.totalTokens).toBe(225);
   });
 
+  it("should include cache/reasoning token usage from GenAI spans", async () => {
+    const span = createMockSpan("gen_ai.chat", {
+      attributes: {
+        "ai.model.id": undefined,
+        "ai.model.provider": undefined,
+        "gen_ai.system": "anthropic",
+        "gen_ai.request.model": "claude-3-5-sonnet",
+        "gen_ai.usage.input_tokens": 150,
+        "gen_ai.usage.output_tokens": 75,
+        "gen_ai.usage.cache_read.input_tokens": 100,
+        "gen_ai.usage.cache_creation.input_tokens": 20,
+        "gen_ai.usage.reasoning.output_tokens": 15
+      }
+    });
+
+    processor.onEnd(span);
+    await processor.forceFlush();
+
+    const event = (mockStore.saveEvent as any).mock.calls[0][0];
+
+    expect(event.payload.usage.cachedTokens).toBe(100);
+    expect(event.payload.usage.cacheCreationTokens).toBe(20);
+    expect(event.payload.usage.reasoningTokens).toBe(15);
+  });
+
   it("should ignore non-AI spans", async () => {
     const span = createMockSpan("http.get");
 

package/test/utils.test.ts

@@ -106,6 +106,38 @@ describe("otel utils", () => {
       expect(attrs.genAiOutputTokens).toBe(25);
     });
 
+    it("should extract OTel GenAI cache/reasoning token attributes", () => {
+      const span = {
+        attributes: {
+          "gen_ai.usage.input_tokens": 100,
+          "gen_ai.usage.output_tokens": 50,
+          "gen_ai.usage.cache_read.input_tokens": 75,
+          "gen_ai.usage.cache_creation.input_tokens": 25,
+          "gen_ai.usage.reasoning.output_tokens": 10
+        }
+      } as any;
+
+      const attrs = extractAIAttributes(span);
+      expect(attrs.genAiCacheReadInputTokens).toBe(75);
+      expect(attrs.genAiCacheCreationInputTokens).toBe(25);
+      expect(attrs.genAiReasoningOutputTokens).toBe(10);
+    });
+
+    it("should extract Vercel AI SDK cache/reasoning token attributes", () => {
+      const span = {
+        attributes: {
+          "ai.usage.promptTokens": 100,
+          "ai.usage.completionTokens": 50,
+          "ai.usage.cachedInputTokens": 60,
+          "ai.usage.reasoningTokens": 12
+        }
+      } as any;
+
+      const attrs = extractAIAttributes(span);
+      expect(attrs.cachedInputTokens).toBe(60);
+      expect(attrs.reasoningTokens).toBe(12);
+    });
+
     it("should extract custom metadata", () => {
       const span = {
         attributes: {