feat(core)!: Remove streamGenAiSpans flag

Removes the `streamGenAiSpans` client option. The behavior it gated
(extracting gen_ai spans from transactions into a v2 span envelope
container) was already the default and is now unconditional — there is
no longer a way to opt out.

For static transactions, `extractGenAiSpansFromEvent` always pulls
gen_ai spans into a v2 span container. For full span streaming
(`traceLifecycle: 'stream'`), the existing `hasSpanStreamingEnabled`
guard still short-circuits so spans are not extracted twice.

Because gen_ai spans now always take the v2 path (not subject to the
transaction payload-size limits that truncation works around),
`shouldEnableTruncation` simplifies to `!getClient()`. The
`enableTruncation` escape hatch is intentionally left in place for a
follow-up.

Test suites that set `streamGenAiSpans: false` to assert on
`transaction.spans` are migrated to assert on the v2 span container;
redundant `streamGenAiSpans: true` config lines are removed.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>

Author: nicohrubec

dev-packages/cloudflare-integration-tests/suites/tracing/anthropic-ai/index.ts

@@ -27,7 +27,6 @@ export default Sentry.withSentry(
   (env: Env) => ({
     dsn: env.SENTRY_DSN,
     tracesSampleRate: 1.0,
-    streamGenAiSpans: true,
   }),
   {
     async fetch(_request, _env, _ctx) {

dev-packages/cloudflare-integration-tests/suites/tracing/google-genai/index.ts

@@ -45,7 +45,6 @@ export default Sentry.withSentry(
   (env: Env) => ({
     dsn: env.SENTRY_DSN,
     tracesSampleRate: 1.0,
-    streamGenAiSpans: true,
   }),
   {
     async fetch(_request, _env, _ctx) {

dev-packages/cloudflare-integration-tests/suites/tracing/langchain/index.ts

@@ -30,7 +30,6 @@ export default Sentry.withSentry(
   (env: Env) => ({
     dsn: env.SENTRY_DSN,
     tracesSampleRate: 1.0,
-    streamGenAiSpans: true,
   }),
   {
     async fetch(_request, _env, _ctx) {

dev-packages/cloudflare-integration-tests/suites/tracing/langgraph/index.ts

@@ -10,7 +10,6 @@ export default Sentry.withSentry(
     dsn: env.SENTRY_DSN,
     tracesSampleRate: 1.0,
     dataCollection: { genAI: { inputs: true, outputs: true } },
-    streamGenAiSpans: true,
   }),
   {
     async fetch(_request, _env, _ctx) {

dev-packages/cloudflare-integration-tests/suites/tracing/openai/index.ts

@@ -32,7 +32,6 @@ export default Sentry.withSentry(
   (env: Env) => ({
     dsn: env.SENTRY_DSN,
     tracesSampleRate: 1.0,
-    streamGenAiSpans: true,
   }),
   {
     async fetch(_request, _env, _ctx) {

dev-packages/cloudflare-integration-tests/suites/tracing/workers-ai/index.ts

@@ -12,9 +12,6 @@ export default Sentry.withSentry(
   (env: Env) => ({
     dsn: env.SENTRY_DSN,
     tracesSampleRate: 1.0,
-    // Keep gen_ai spans embedded in the transaction (instead of streamed as a
-    // separate envelope container) so they can be asserted on `transaction.spans`.
-    streamGenAiSpans: false,
   }),
   {
     async fetch(request) {

dev-packages/cloudflare-integration-tests/suites/tracing/workers-ai/test.ts

@@ -22,40 +22,29 @@ it('traces a basic Workers AI text generation request', async ({ signal }) => {
     .ignore('event')
     .expect(envelope => {
       const transactionEvent = envelope[1]?.[0]?.[1] as any;
+      expect(transactionEvent.transaction).toBe('GET /');
 
-      // The transaction event is framework-generated and carries non-deterministic fields
-      // (random ports, ids, timestamps, sdk version), so we assert the stable subset.
-      expect(transactionEvent).toEqual(
+      const container = envelope[1]?.[1]?.[1] as any;
+      expect(container).toBeDefined();
+      expect(container.items).toHaveLength(1);
+
+      expect(container.items[0]).toEqual(
         expect.objectContaining({
-          type: 'transaction',
-          transaction: 'GET /',
-          transaction_info: { source: 'route' },
-          contexts: expect.objectContaining({
-            trace: expect.objectContaining({
-              op: 'http.server',
-              origin: 'auto.http.cloudflare',
-              status: 'ok',
-            }),
-          }),
-          spans: [
-            expect.objectContaining({
-              description: 'chat @cf/meta/llama-3.1-8b-instruct',
-              op: 'gen_ai.chat',
-              origin: 'auto.ai.cloudflare.workers_ai',
-              data: {
-                'sentry.origin': 'auto.ai.cloudflare.workers_ai',
-                'sentry.op': 'gen_ai.chat',
-                [GEN_AI_PROVIDER_NAME]: 'cloudflare.workers_ai',
-                [GEN_AI_OPERATION_NAME_ATTRIBUTE]: 'chat',
-                [GEN_AI_REQUEST_MODEL_ATTRIBUTE]: '@cf/meta/llama-3.1-8b-instruct',
-                [GEN_AI_REQUEST_TEMPERATURE_ATTRIBUTE]: 0.7,
-                [GEN_AI_REQUEST_MAX_TOKENS_ATTRIBUTE]: 100,
-                [GEN_AI_USAGE_INPUT_TOKENS_ATTRIBUTE]: 12,
-                [GEN_AI_USAGE_OUTPUT_TOKENS_ATTRIBUTE]: 7,
-                [GEN_AI_USAGE_TOTAL_TOKENS_ATTRIBUTE]: 19,
-              },
-            }),
-          ],
+          name: 'chat @cf/meta/llama-3.1-8b-instruct',
+          status: 'ok',
+          is_segment: false,
+          attributes: {
+            'sentry.origin': { value: 'auto.ai.cloudflare.workers_ai', type: 'string' },
+            'sentry.op': { value: 'gen_ai.chat', type: 'string' },
+            [GEN_AI_PROVIDER_NAME]: { value: 'cloudflare.workers_ai', type: 'string' },
+            [GEN_AI_OPERATION_NAME_ATTRIBUTE]: { value: 'chat', type: 'string' },
+            [GEN_AI_REQUEST_MODEL_ATTRIBUTE]: { value: '@cf/meta/llama-3.1-8b-instruct', type: 'string' },
+            [GEN_AI_REQUEST_TEMPERATURE_ATTRIBUTE]: { value: 0.7, type: 'double' },
+            [GEN_AI_REQUEST_MAX_TOKENS_ATTRIBUTE]: { value: 100, type: 'integer' },
+            [GEN_AI_USAGE_INPUT_TOKENS_ATTRIBUTE]: { value: 12, type: 'integer' },
+            [GEN_AI_USAGE_OUTPUT_TOKENS_ATTRIBUTE]: { value: 7, type: 'integer' },
+            [GEN_AI_USAGE_TOTAL_TOKENS_ATTRIBUTE]: { value: 19, type: 'integer' },
+          },
         }),
       );
     })
@@ -69,38 +58,29 @@ it('traces a streaming Workers AI text generation request', async ({ signal }) =
     .ignore('event')
     .expect(envelope => {
       const transactionEvent = envelope[1]?.[0]?.[1] as any;
+      expect(transactionEvent.transaction).toBe('GET /stream');
 
-      expect(transactionEvent).toEqual(
+      const container = envelope[1]?.[1]?.[1] as any;
+      expect(container).toBeDefined();
+      expect(container.items).toHaveLength(1);
+
+      expect(container.items[0]).toEqual(
         expect.objectContaining({
-          type: 'transaction',
-          transaction: 'GET /stream',
-          transaction_info: { source: 'url' },
-          contexts: expect.objectContaining({
-            trace: expect.objectContaining({
-              op: 'http.server',
-              origin: 'auto.http.cloudflare',
-              status: 'ok',
-            }),
-          }),
-          spans: [
-            expect.objectContaining({
-              description: 'chat @cf/meta/llama-3.1-8b-instruct',
-              op: 'gen_ai.chat',
-              origin: 'auto.ai.cloudflare.workers_ai',
-              data: {
-                'sentry.origin': 'auto.ai.cloudflare.workers_ai',
-                'sentry.op': 'gen_ai.chat',
-                [GEN_AI_PROVIDER_NAME]: 'cloudflare.workers_ai',
-                [GEN_AI_OPERATION_NAME_ATTRIBUTE]: 'chat',
-                [GEN_AI_REQUEST_MODEL_ATTRIBUTE]: '@cf/meta/llama-3.1-8b-instruct',
-                [GEN_AI_REQUEST_STREAM_ATTRIBUTE]: true,
-                [GEN_AI_RESPONSE_STREAMING_ATTRIBUTE]: true,
-                [GEN_AI_USAGE_INPUT_TOKENS_ATTRIBUTE]: 12,
-                [GEN_AI_USAGE_OUTPUT_TOKENS_ATTRIBUTE]: 7,
-                [GEN_AI_USAGE_TOTAL_TOKENS_ATTRIBUTE]: 19,
-              },
-            }),
-          ],
+          name: 'chat @cf/meta/llama-3.1-8b-instruct',
+          status: 'ok',
+          is_segment: false,
+          attributes: {
+            'sentry.origin': { value: 'auto.ai.cloudflare.workers_ai', type: 'string' },
+            'sentry.op': { value: 'gen_ai.chat', type: 'string' },
+            [GEN_AI_PROVIDER_NAME]: { value: 'cloudflare.workers_ai', type: 'string' },
+            [GEN_AI_OPERATION_NAME_ATTRIBUTE]: { value: 'chat', type: 'string' },
+            [GEN_AI_REQUEST_MODEL_ATTRIBUTE]: { value: '@cf/meta/llama-3.1-8b-instruct', type: 'string' },
+            [GEN_AI_REQUEST_STREAM_ATTRIBUTE]: { value: true, type: 'boolean' },
+            [GEN_AI_RESPONSE_STREAMING_ATTRIBUTE]: { value: true, type: 'boolean' },
+            [GEN_AI_USAGE_INPUT_TOKENS_ATTRIBUTE]: { value: 12, type: 'integer' },
+            [GEN_AI_USAGE_OUTPUT_TOKENS_ATTRIBUTE]: { value: 7, type: 'integer' },
+            [GEN_AI_USAGE_TOTAL_TOKENS_ATTRIBUTE]: { value: 19, type: 'integer' },
+          },
         }),
       );
     })

dev-packages/e2e-tests/test-applications/cloudflare-vercelai-v7-als/src/index.ts

@@ -7,7 +7,6 @@ export default Sentry.withSentry(
     dsn: env.E2E_TEST_DSN,
     environment: 'qa',
     tunnel: 'http://localhost:3031/',
-    streamGenAiSpans: false,
     tracesSampleRate: 1.0,
     integrations: [Sentry.vercelAIIntegration()],
   }),

dev-packages/e2e-tests/test-applications/cloudflare-vercelai-v7-als/tests/index.test.ts

@@ -1,23 +1,33 @@
 import { expect, test } from '@playwright/test';
-import { waitForTransaction } from '@sentry-internal/test-utils';
+import { getSpanOp, waitForRequest } from '@sentry-internal/test-utils';
 
 test('does not capture Vercel AI v7 spans without nodejs_compat', async ({ baseURL }) => {
-  const transactionPromise = waitForTransaction('cloudflare-vercelai-v7-als', txn => {
-    return txn.transaction === 'GET /generate';
+  // The transaction envelope also carries any extracted gen_ai spans as a span v2 container item,
+  // so we wait for the whole envelope and assert neither place contains AI spans.
+  const envelopePromise = waitForRequest('cloudflare-vercelai-v7-als', ({ envelope }) => {
+    const transactionItem = envelope[1].find(([header]) => header.type === 'transaction');
+    return (transactionItem?.[1] as any)?.transaction === 'GET /generate';
   });
 
   const response = await fetch(`${baseURL}/generate`);
   expect(response.status).toBe(200);
 
-  const transaction = await transactionPromise;
+  const { envelope } = await envelopePromise;
 
+  const transaction = envelope[1].find(([header]) => header.type === 'transaction')?.[1] as any;
   expect(transaction.transaction).toBe('GET /generate');
   expect(transaction.contexts?.trace?.op).toBe('http.server');
 
-  // v7 uses diagnostics_channel which is not available with nodejs_als,
-  // so no AI spans should be present.
-  const aiSpans = (transaction.spans || []).filter(
+  // v7 uses diagnostics_channel which is not available with nodejs_als, so no AI spans should be
+  // present — neither embedded in the transaction nor streamed as a span v2 container item.
+  const embeddedAiSpans = (transaction.spans || []).filter(
     (span: any) => span.op?.startsWith('gen_ai.') || span.description?.includes('generateText'),
   );
-  expect(aiSpans).toHaveLength(0);
+  expect(embeddedAiSpans).toHaveLength(0);
+
+  const streamedGenAiSpans = envelope[1]
+    .filter(([header]) => header.type === 'span')
+    .flatMap(([, payload]) => (payload as any).items ?? [])
+    .filter((span: any) => getSpanOp(span)?.startsWith('gen_ai.'));
+  expect(streamedGenAiSpans).toHaveLength(0);
 });

dev-packages/e2e-tests/test-applications/cloudflare-vercelai-v7-compat/src/index.ts

@@ -8,7 +8,6 @@ export default Sentry.withSentry(
     environment: 'qa',
     tunnel: 'http://localhost:3031/',
     tracesSampleRate: 1.0,
-    streamGenAiSpans: false,
     integrations: [Sentry.vercelAIIntegration()],
   }),
   {