fix(core): Instrument invoke_agent root span, and support Vercel ai v5 (#17395)

Main changes are: 
1. Instrument root `invoke_agent` span: This addresses that transaction
event.spans only includes child spans, ensuring proper correlation
beyond just children.
2. Refactor `vercelai` integration layout in core: Restructured the
vercelai folder to mirror our other AI agent integrations (split
constants/utils/index) for clarity and easier maintenance.
3. Link errors in Vercel AI v5 (metadata-based): Added support to
capture and link tool errors that are returned as result metadata (not
thrown). We correlate tool-call IDs to spans and set the trace/span
context when capturing the error, so errors are attached to the right
tool-call/trace even without a thrown exception.
4. Added tests to test vercel ai v5

Author: RulaKhaled

dev-packages/node-integration-tests/suites/tracing/vercelai/test.ts

@@ -197,73 +197,6 @@ describe('Vercel AI integration', () => {
     ]),
   };
 
-  // Todo: Add missing attribute spans for v5
-  // Right now only second span is recorded as it's manually opted in via explicit telemetry option
-  const EXPECTED_TRANSACTION_DEFAULT_PII_FALSE_V5 = {
-    transaction: 'main',
-    spans: expect.arrayContaining([
-      expect.objectContaining({
-        data: {
-          'vercel.ai.model.id': 'mock-model-id',
-          'vercel.ai.model.provider': 'mock-provider',
-          'vercel.ai.operationId': 'ai.generateText',
-          'vercel.ai.pipeline.name': 'generateText',
-          'vercel.ai.prompt': '{"prompt":"Where is the second span?"}',
-          'vercel.ai.response.finishReason': 'stop',
-          'gen_ai.response.text': expect.any(String),
-          'vercel.ai.settings.maxRetries': 2,
-          // 'vercel.ai.settings.maxSteps': 1,
-          'vercel.ai.streaming': false,
-          'gen_ai.prompt': '{"prompt":"Where is the second span?"}',
-          'gen_ai.response.model': 'mock-model-id',
-          'gen_ai.usage.input_tokens': 10,
-          'gen_ai.usage.output_tokens': 20,
-          'gen_ai.usage.total_tokens': 30,
-          'operation.name': 'ai.generateText',
-          'sentry.op': 'gen_ai.invoke_agent',
-          'sentry.origin': 'auto.vercelai.otel',
-        },
-        description: 'generateText',
-        op: 'gen_ai.invoke_agent',
-        origin: 'auto.vercelai.otel',
-        status: 'ok',
-      }),
-      // doGenerate
-      expect.objectContaining({
-        data: {
-          'sentry.origin': 'auto.vercelai.otel',
-          'sentry.op': 'gen_ai.generate_text',
-          'operation.name': 'ai.generateText.doGenerate',
-          'vercel.ai.operationId': 'ai.generateText.doGenerate',
-          'vercel.ai.model.provider': 'mock-provider',
-          'vercel.ai.model.id': 'mock-model-id',
-          'vercel.ai.settings.maxRetries': 2,
-          'gen_ai.system': 'mock-provider',
-          'gen_ai.request.model': 'mock-model-id',
-          'vercel.ai.pipeline.name': 'generateText.doGenerate',
-          'vercel.ai.streaming': false,
-          'vercel.ai.response.finishReason': 'stop',
-          'vercel.ai.response.model': 'mock-model-id',
-          'vercel.ai.response.id': expect.any(String),
-          'gen_ai.response.text': 'Second span here!',
-          'vercel.ai.response.timestamp': expect.any(String),
-          // 'vercel.ai.prompt.format': expect.any(String),
-          'gen_ai.request.messages': expect.any(String),
-          'gen_ai.response.finish_reasons': ['stop'],
-          'gen_ai.usage.input_tokens': 10,
-          'gen_ai.usage.output_tokens': 20,
-          'gen_ai.response.id': expect.any(String),
-          'gen_ai.response.model': 'mock-model-id',
-          'gen_ai.usage.total_tokens': 30,
-        },
-        description: 'generate_text mock-model-id',
-        op: 'gen_ai.generate_text',
-        origin: 'auto.vercelai.otel',
-        status: 'ok',
-      }),
-    ]),
-  };
-
   const EXPECTED_TRANSACTION_DEFAULT_PII_TRUE = {
     transaction: 'main',
     spans: expect.arrayContaining([
@@ -605,23 +538,6 @@ describe('Vercel AI integration', () => {
     });
   });
 
-  // Test with specific Vercel AI v5 version
-  createEsmAndCjsTests(
-    __dirname,
-    'scenario-v5.mjs',
-    'instrument.mjs',
-    (createRunner, test) => {
-      test('creates ai related spans with v5', async () => {
-        await createRunner().expect({ transaction: EXPECTED_TRANSACTION_DEFAULT_PII_FALSE_V5 }).start().completed();
-      });
-    },
-    {
-      additionalDependencies: {
-        ai: '^5.0.0',
-      },
-    },
-  );
-
   createEsmAndCjsTests(__dirname, 'scenario-error-in-tool-express.mjs', 'instrument.mjs', (createRunner, test) => {
     test('captures error in tool in express server', async () => {
       const expectedTransaction = {

dev-packages/node-integration-tests/suites/tracing/vercelai/v5/instrument-with-pii.mjs

@@ -0,0 +1,11 @@
+import * as Sentry from '@sentry/node';
+import { loggingTransport } from '@sentry-internal/node-integration-tests';
+
+Sentry.init({
+  dsn: 'https://public@dsn.ingest.sentry.io/1337',
+  release: '1.0',
+  tracesSampleRate: 1.0,
+  sendDefaultPii: true,
+  transport: loggingTransport,
+  integrations: [Sentry.vercelAIIntegration()],
+});

dev-packages/node-integration-tests/suites/tracing/vercelai/v5/instrument.mjs

@@ -0,0 +1,10 @@
+import * as Sentry from '@sentry/node';
+import { loggingTransport } from '@sentry-internal/node-integration-tests';
+
+Sentry.init({
+  dsn: 'https://public@dsn.ingest.sentry.io/1337',
+  release: '1.0',
+  tracesSampleRate: 1.0,
+  transport: loggingTransport,
+  integrations: [Sentry.vercelAIIntegration()],
+});

dev-packages/node-integration-tests/suites/tracing/vercelai/v5/scenario-error-in-tool.mjs

@@ -0,0 +1,36 @@
+import * as Sentry from '@sentry/node';
+import { generateText, tool } from 'ai';
+import { MockLanguageModelV2 } from 'ai/test';
+import { z } from 'zod';
+
+async function run() {
+  await Sentry.startSpan({ op: 'function', name: 'main' }, async () => {
+    await generateText({
+      model: new MockLanguageModelV2({
+        doGenerate: async () => ({
+          finishReason: 'tool-calls',
+          usage: { inputTokens: 15, outputTokens: 25, totalTokens: 40 },
+          content: [
+            {
+              type: 'tool-call',
+              toolCallId: 'call-1',
+              toolName: 'getWeather',
+              input: JSON.stringify({ location: 'San Francisco' }),
+            },
+          ],
+        }),
+      }),
+      tools: {
+        getWeather: tool({
+          inputSchema: z.object({ location: z.string() }),
+          execute: async () => {
+            throw new Error('Error in tool');
+          },
+        }),
+      },
+      prompt: 'What is the weather in San Francisco?',
+    });
+  });
+}
+
+run();

dev-packages/node-integration-tests/suites/tracing/vercelai/scenario-v5.mjs renamed to dev-packages/node-integration-tests/suites/tracing/vercelai/v5/scenario.mjs

@@ -1,5 +1,5 @@
 import * as Sentry from '@sentry/node';
-import { generateText } from 'ai';
+import { generateText, tool } from 'ai';
 import { MockLanguageModelV2 } from 'ai/test';
 import { z } from 'zod';
 
@@ -35,24 +35,21 @@ async function run() {
         doGenerate: async () => ({
           finishReason: 'tool-calls',
           usage: { inputTokens: 15, outputTokens: 25, totalTokens: 40 },
-          content: [{ type: 'text', text: 'Tool call completed!' }],
-          toolCalls: [
+          content: [
             {
-              toolCallType: 'function',
+              type: 'tool-call',
               toolCallId: 'call-1',
               toolName: 'getWeather',
-              args: '{ "location": "San Francisco" }',
+              input: JSON.stringify({ location: 'San Francisco' }),
             },
           ],
         }),
       }),
       tools: {
-        getWeather: {
-          parameters: z.object({ location: z.string() }),
-          execute: async args => {
-            return `Weather in ${args.location}: Sunny, 72°F`;
-          },
-        },
+        getWeather: tool({
+          inputSchema: z.object({ location: z.string() }),
+          execute: async ({ location }) => `Weather in ${location}: Sunny, 72°F`,
+        }),
       },
       prompt: 'What is the weather in San Francisco?',
     });