feat(core,node): Add instrumentation for GoogleGenerativeAI

[RulaKhaled]

This PR implements instrumentation for the Google GenerativeAI SDK, adding automatic tracing for both direct model calls (models.generateContent) and conversational chat sessions (chats.create + sendMessage). The implementation follows Sentry's AI Agents Manual Instrumentation conventions and includes integration tests.

Usage:

import * as Sentry from '@sentry/node';

Sentry.init({
  integrations: [
    Sentry.googleGenAIIntegration({
      recordInputs: true,   // Record prompts/messages
      recordOutputs: true,  // Record AI responses
    }),
  ],
});

// Chat Google GenAI calls are now automatically instrumented

[github-actions]

size-limit report 📦

Path	Size	% Change	Change
@sentry/browser	24.22 kB	-	-
@sentry/browser - with treeshaking flags	22.74 kB	-	-
@sentry/browser (incl. Tracing)	40.26 kB	-	-
@sentry/browser (incl. Tracing, Replay)	78.63 kB	-	-
@sentry/browser (incl. Tracing, Replay) - with treeshaking flags	68.29 kB	-	-
@sentry/browser (incl. Tracing, Replay with Canvas)	83.29 kB	-	-
@sentry/browser (incl. Tracing, Replay, Feedback)	95.52 kB	-	-
@sentry/browser (incl. Feedback)	40.94 kB	-	-
@sentry/browser (incl. sendFeedback)	28.88 kB	-	-
@sentry/browser (incl. FeedbackAsync)	33.78 kB	-	-
@sentry/react	25.93 kB	-	-
@sentry/react (incl. Tracing)	42.26 kB	-	-
@sentry/vue	28.71 kB	-	-
@sentry/vue (incl. Tracing)	42.07 kB	-	-
@sentry/svelte	24.24 kB	-	-
CDN Bundle	25.75 kB	-	-
CDN Bundle (incl. Tracing)	40.07 kB	-	-
CDN Bundle (incl. Tracing, Replay)	76.3 kB	-	-
CDN Bundle (incl. Tracing, Replay, Feedback)	81.8 kB	-	-
CDN Bundle - uncompressed	75.23 kB	-	-
CDN Bundle (incl. Tracing) - uncompressed	118.65 kB	-	-
CDN Bundle (incl. Tracing, Replay) - uncompressed	233.78 kB	-	-
CDN Bundle (incl. Tracing, Replay, Feedback) - uncompressed	246.54 kB	-	-
@sentry/nextjs (client)	44.26 kB	-	-
@sentry/sveltekit (client)	40.69 kB	-	-
@sentry/node-core	49.96 kB	-0.01%	-1 B 🔽
@sentry/node	152.25 kB	+0.66%	+993 B 🔺
@sentry/node - without tracing	91.86 kB	-0.01%	-2 B 🔽
@sentry/aws-serverless	105.31 kB	-	-

View base workflow run

[github-actions]

node-overhead report 🧳

Note: This is a synthetic benchmark with a minimal express app and does not necessarily reflect the real-world performance impact in an application.

Scenario	Requests/s	% of Baseline	Prev. Requests/s	Change %
GET Baseline	8,636	-	9,143	-6%
GET With Sentry	1,373	16%	1,394	-2%
GET With Sentry (error only)	5,927	69%	6,056	-2%
POST Baseline	1,181	-	1,199	-2%
POST With Sentry	549	46%	531	+3%
POST With Sentry (error only)	1,048	89%	1,065	-2%
MYSQL Baseline	3,270	-	3,318	-1%
MYSQL With Sentry	486	15%	453	+7%
MYSQL With Sentry (error only)	2,674	82%	2,670	+0%

View base workflow run

[mydea]

Suggested change

integrations: [Sentry.googleGenAIIntegration()],

this is added by default, right? So no need to add this here :)

[mydea]

Suggested change

integrations: [Sentry.googleGenAIIntegration()],

[mydea]

do we actually need this, when we do not lookup the integration on the client anymore? :)

[mydea]

👍 nice util!

[mydea]

let's add a comment here why this is necessary, for future reference!

[mydea]

Suggested change

	return;
	return exports;

Do we not need to return the exports here if we do not patch it? 🤔 or is this handled if this is undefined and we do nothing then? just double checking 😅

[RulaKhaled]

Nah i think you're correct

[mydea]

left some small comments but overall this looks very good now 🚀