ref(core)!: Rename gen_ai.system to gen_ai.provider.name

Emit the `gen_ai.provider.name` attribute from `@sentry/conventions` instead
of the deprecated `gen_ai.system` alias across all AI integrations (OpenAI,
Anthropic, Google GenAI, LangChain, Bedrock, Vercel AI). The OTel-based
Vercel AI processor now renames the SDK-native `gen_ai.system` attribute to
match.

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

Author: nicohrubec

MIGRATION.md

@@ -252,6 +252,7 @@ Affected SDKs: All SDKs.
 - `network.*` span attributes were aligned across SDKs.
 - Legacy messaging (`messaging.*`) and database (`db.statement`, …) span attributes on the AMQP and Redis instrumentations were replaced by their current semantic-convention equivalents.
 - The gen_ai cache token attributes `gen_ai.usage.cache_creation_input_tokens` and `gen_ai.usage.cache_read_input_tokens` were renamed to `gen_ai.usage.cache_creation.input_tokens` and `gen_ai.usage.cache_read.input_tokens`.
+- The `gen_ai.system` span attribute was renamed to `gen_ai.provider.name` across all AI integrations.
 - Span attributes now use the shared `@sentry/conventions` package under the hood.
 
 If you reference these attributes in custom instrumentation, `beforeSendSpan`, dashboards, or alerts, update them to the new names.

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

@@ -2,13 +2,13 @@ import { expect, it } from 'vitest';
 import {
   GEN_AI_INPUT_MESSAGES,
   GEN_AI_OPERATION_NAME,
+  GEN_AI_PROVIDER_NAME,
   GEN_AI_REQUEST_MAX_TOKENS,
   GEN_AI_REQUEST_MODEL,
   GEN_AI_REQUEST_TEMPERATURE,
   GEN_AI_RESPONSE_ID,
   GEN_AI_RESPONSE_MODEL,
   GEN_AI_RESPONSE_TEXT,
-  GEN_AI_SYSTEM,
   GEN_AI_USAGE_INPUT_TOKENS,
   GEN_AI_USAGE_OUTPUT_TOKENS,
   GEN_AI_USAGE_TOTAL_TOKENS,
@@ -42,7 +42,7 @@ it('traces a basic message creation request with the anthropic SDK', async ({ si
         attributes: {
           'sentry.origin': { value: 'auto.ai.anthropic', type: 'string' },
           'sentry.op': { value: 'gen_ai.chat', type: 'string' },
-          [GEN_AI_SYSTEM]: { value: 'anthropic', type: 'string' },
+          [GEN_AI_PROVIDER_NAME]: { value: 'anthropic', type: 'string' },
           [GEN_AI_OPERATION_NAME]: { value: 'chat', type: 'string' },
           [GEN_AI_REQUEST_MODEL]: { value: 'claude-3-haiku-20240307', type: 'string' },
           [GEN_AI_REQUEST_TEMPERATURE]: { value: 0.7, type: 'double' },

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

@@ -4,12 +4,12 @@ import {
   GEN_AI_EMBEDDINGS_INPUT,
   GEN_AI_INPUT_MESSAGES,
   GEN_AI_OPERATION_NAME,
+  GEN_AI_PROVIDER_NAME,
   GEN_AI_REQUEST_MAX_TOKENS,
   GEN_AI_REQUEST_MODEL,
   GEN_AI_REQUEST_TEMPERATURE,
   GEN_AI_REQUEST_TOP_P,
   GEN_AI_RESPONSE_TEXT,
-  GEN_AI_SYSTEM,
   GEN_AI_USAGE_INPUT_TOKENS,
   GEN_AI_USAGE_OUTPUT_TOKENS,
   GEN_AI_USAGE_TOTAL_TOKENS,
@@ -46,7 +46,7 @@ it('traces Google GenAI chat, generateContent, and embedContent calls', async ({
         attributes: {
           'sentry.origin': { value: 'auto.ai.google_genai', type: 'string' },
           'sentry.op': { value: 'gen_ai.chat', type: 'string' },
-          [GEN_AI_SYSTEM]: { value: 'google_genai', type: 'string' },
+          [GEN_AI_PROVIDER_NAME]: { value: 'google_genai', type: 'string' },
           [GEN_AI_OPERATION_NAME]: { value: 'chat', type: 'string' },
           [GEN_AI_REQUEST_MODEL]: { value: 'gemini-1.5-pro', type: 'string' },
           // collect LLM input and outputs (default true)
@@ -70,7 +70,7 @@ it('traces Google GenAI chat, generateContent, and embedContent calls', async ({
         attributes: {
           'sentry.origin': { value: 'auto.ai.google_genai', type: 'string' },
           'sentry.op': { value: 'gen_ai.generate_content', type: 'string' },
-          [GEN_AI_SYSTEM]: { value: 'google_genai', type: 'string' },
+          [GEN_AI_PROVIDER_NAME]: { value: 'google_genai', type: 'string' },
           [GEN_AI_OPERATION_NAME]: { value: 'generate_content', type: 'string' },
           [GEN_AI_REQUEST_MODEL]: { value: 'gemini-1.5-flash', type: 'string' },
           [GEN_AI_REQUEST_TEMPERATURE]: { value: 0.7, type: 'double' },
@@ -99,7 +99,7 @@ it('traces Google GenAI chat, generateContent, and embedContent calls', async ({
         attributes: {
           'sentry.origin': { value: 'auto.ai.google_genai', type: 'string' },
           'sentry.op': { value: 'gen_ai.embeddings', type: 'string' },
-          [GEN_AI_SYSTEM]: { value: 'google_genai', type: 'string' },
+          [GEN_AI_PROVIDER_NAME]: { value: 'google_genai', type: 'string' },
           [GEN_AI_OPERATION_NAME]: { value: 'embeddings', type: 'string' },
           [GEN_AI_REQUEST_MODEL]: { value: 'text-embedding-004', type: 'string' },
           [GEN_AI_EMBEDDINGS_INPUT]: { value: 'Hello world', type: 'string' },

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

@@ -1,13 +1,13 @@
 import { expect, it } from 'vitest';
 import {
   GEN_AI_OPERATION_NAME,
+  GEN_AI_PROVIDER_NAME,
   GEN_AI_REQUEST_MAX_TOKENS,
   GEN_AI_REQUEST_MODEL,
   GEN_AI_REQUEST_TEMPERATURE,
   GEN_AI_RESPONSE_FINISH_REASONS,
   GEN_AI_RESPONSE_ID,
   GEN_AI_RESPONSE_MODEL,
-  GEN_AI_SYSTEM,
   GEN_AI_USAGE_INPUT_TOKENS,
   GEN_AI_USAGE_OUTPUT_TOKENS,
   GEN_AI_USAGE_TOTAL_TOKENS,
@@ -42,7 +42,7 @@ it('traces a LangChain chat model invocation', async ({ signal }) => {
         attributes: {
           'sentry.origin': { value: 'auto.ai.langchain', type: 'string' },
           'sentry.op': { value: 'gen_ai.chat', type: 'string' },
-          [GEN_AI_SYSTEM]: { value: 'openai', type: 'string' },
+          [GEN_AI_PROVIDER_NAME]: { value: 'openai', type: 'string' },
           [GEN_AI_OPERATION_NAME]: { value: 'chat', type: 'string' },
           [GEN_AI_REQUEST_MODEL]: { value: 'gpt-3.5-turbo', type: 'string' },
           [GEN_AI_REQUEST_TEMPERATURE]: { value: 0.7, type: 'double' },

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

@@ -2,12 +2,12 @@ import { expect, it } from 'vitest';
 import {
   GEN_AI_INPUT_MESSAGES,
   GEN_AI_OPERATION_NAME,
+  GEN_AI_PROVIDER_NAME,
   GEN_AI_REQUEST_MODEL,
   GEN_AI_REQUEST_TEMPERATURE,
   GEN_AI_RESPONSE_FINISH_REASONS,
   GEN_AI_RESPONSE_ID,
   GEN_AI_RESPONSE_MODEL,
-  GEN_AI_SYSTEM,
   GEN_AI_SYSTEM_INSTRUCTIONS,
   GEN_AI_USAGE_INPUT_TOKENS,
   GEN_AI_USAGE_OUTPUT_TOKENS,
@@ -42,7 +42,7 @@ it('traces a basic chat completion request with the openai SDK', async ({ signal
         attributes: {
           'sentry.origin': { value: 'auto.ai.openai', type: 'string' },
           'sentry.op': { value: 'gen_ai.chat', type: 'string' },
-          [GEN_AI_SYSTEM]: { value: 'openai', type: 'string' },
+          [GEN_AI_PROVIDER_NAME]: { value: 'openai', type: 'string' },
           [GEN_AI_OPERATION_NAME]: { value: 'chat', type: 'string' },
           [GEN_AI_REQUEST_MODEL]: { value: 'gpt-3.5-turbo', type: 'string' },
           [GEN_AI_REQUEST_TEMPERATURE]: { value: 0.7, type: 'double' },

dev-packages/deno-integration-tests/suites/orchestrion-anthropic/test.ts

@@ -51,7 +51,7 @@ Deno.test('anthropic instrumentation: orchestrion @anthropic-ai/sdk:chat channel
   const aiSpan = parent.spans?.find(s => s.op === 'gen_ai.chat');
   assertExists(aiSpan, `expected a gen_ai.chat child span, got ops: ${parent.spans?.map(s => s.op).join(', ')}`);
   assertEquals(aiSpan!.description, 'chat claude-3-5-sonnet-latest');
-  assertEquals(aiSpan!.data?.['gen_ai.system'], 'anthropic');
+  assertEquals(aiSpan!.data?.['gen_ai.provider.name'], 'anthropic');
   assertEquals(aiSpan!.data?.['gen_ai.operation.name'], 'chat');
   assertEquals(aiSpan!.data?.['gen_ai.request.model'], 'claude-3-5-sonnet-latest');
   assertEquals(aiSpan!.data?.['gen_ai.response.model'], 'claude-3-5-sonnet-20241022');

dev-packages/deno-integration-tests/suites/orchestrion-google-genai/test.ts

@@ -53,7 +53,7 @@ Deno.test('google-genai instrumentation: orchestrion @google/genai:generate-cont
     `expected a gen_ai.generate_content child span, got ops: ${parent.spans?.map(s => s.op).join(', ')}`,
   );
   assertEquals(aiSpan!.description, 'generate_content gemini-1.5-flash');
-  assertEquals(aiSpan!.data?.['gen_ai.system'], 'google_genai');
+  assertEquals(aiSpan!.data?.['gen_ai.provider.name'], 'google_genai');
   assertEquals(aiSpan!.data?.['gen_ai.operation.name'], 'generate_content');
   assertEquals(aiSpan!.data?.['gen_ai.request.model'], 'gemini-1.5-flash');
   assertEquals(aiSpan!.data?.['gen_ai.response.model'], 'gemini-1.5-flash-002');

dev-packages/deno-integration-tests/suites/orchestrion-langchain/test.ts

@@ -48,7 +48,7 @@ Deno.test('langchain instrumentation: orchestrion @langchain/openai:embedQuery c
   const aiSpan = parent.spans?.find(s => s.op === 'gen_ai.embeddings');
   assertExists(aiSpan, `expected a gen_ai.embeddings child span, got ops: ${parent.spans?.map(s => s.op).join(', ')}`);
   assertEquals(aiSpan!.description, 'embeddings text-embedding-3-small');
-  assertEquals(aiSpan!.data?.['gen_ai.system'], 'openai');
+  assertEquals(aiSpan!.data?.['gen_ai.provider.name'], 'openai');
   assertEquals(aiSpan!.data?.['gen_ai.operation.name'], 'embeddings');
   assertEquals(aiSpan!.data?.['gen_ai.request.model'], 'text-embedding-3-small');
   assertEquals(aiSpan!.data?.['sentry.origin'], 'auto.ai.langchain');

dev-packages/deno-integration-tests/suites/orchestrion-openai/test.ts

@@ -51,7 +51,7 @@ Deno.test('openai instrumentation: orchestrion:openai:chat channel produces a ne
   const aiSpan = parent.spans?.find(s => s.op === 'gen_ai.chat');
   assertExists(aiSpan, `expected a gen_ai.chat child span, got ops: ${parent.spans?.map(s => s.op).join(', ')}`);
   assertEquals(aiSpan!.description, 'chat gpt-4o');
-  assertEquals(aiSpan!.data?.['gen_ai.system'], 'openai');
+  assertEquals(aiSpan!.data?.['gen_ai.provider.name'], 'openai');
   assertEquals(aiSpan!.data?.['gen_ai.operation.name'], 'chat');
   assertEquals(aiSpan!.data?.['gen_ai.request.model'], 'gpt-4o');
   assertEquals(aiSpan!.data?.['gen_ai.response.model'], 'gpt-4o-2024-08-06');

dev-packages/deno-integration-tests/suites/orchestrion-vercel-ai/test.ts

@@ -53,7 +53,7 @@ Deno.test('vercel-ai instrumentation: orchestrion:ai:generateText channel produc
     `expected a gen_ai.invoke_agent child span, got ops: ${parent.spans?.map(s => s.op).join(', ')}`,
   );
   assertEquals(aiSpan!.description, 'invoke_agent');
-  assertEquals(aiSpan!.data?.['gen_ai.system'], 'openai');
+  assertEquals(aiSpan!.data?.['gen_ai.provider.name'], 'openai');
   assertEquals(aiSpan!.data?.['gen_ai.operation.name'], 'invoke_agent');
   assertEquals(aiSpan!.data?.['gen_ai.request.model'], 'gpt-4o');
   assertEquals(aiSpan!.data?.['vercel.ai.operationId'], 'ai.generateText');