feat(aws): Create unified lambda layer for ESM and CJS

[msonnb]

This introduces a new AWS lambda layer that supports both ESM and CJS. Instead of bundling the whole SDK, we install the local NPM package and then prune all not strictly necessary files from node_modules by using @vercel/nft to keep the layer size as small as possible.

closes#16876
closes#16883
closes#16886
closes#16879

[andreiborza]

Great work, just a few minor comments. I think we can give this a spin!

[andreiborza]

🚀

[cursor]

Bug: Inconsistent Path Handling Causes Pruning Failures

The node_modules pruning logic fails due to inconsistent path formatting. Paths from @vercel/nft (based on entrypoints with ./ prefix) and paths from getAllFiles (relative to process.cwd()) do not consistently match, potentially due to leading ./ differences and platform-specific path separators. This causes the file filtering to fail, leading to incorrect pruning (e.g., deleting all files or preserving/deleting the wrong ones).

Additionally, the nodeFileTrace call lacks error handling; if any hardcoded entrypoint files are missing, it will crash the build process.

packages/aws-serverless/scripts/buildLambdaLayer.ts#L66-L78

sentry-javascript/packages/aws-serverless/scripts/buildLambdaLayer.ts

Lines 66 to 78 in 4ad4e49

	asyncfunctionpruneNodeModules(): Promise<void>{
	constentrypoints=[
	'./build/aws/dist-serverless/nodejs/node_modules/@sentry/aws-serverless/build/npm/esm/index.js',
	'./build/aws/dist-serverless/nodejs/node_modules/@sentry/aws-serverless/build/npm/cjs/index.js',
	'./build/aws/dist-serverless/nodejs/node_modules/@sentry/aws-serverless/build/npm/cjs/awslambda-auto.js',
	'./build/aws/dist-serverless/nodejs/node_modules/@sentry/aws-serverless/build/npm/esm/awslambda-auto.js',
	];
	const{ fileList }=awaitnodeFileTrace(entrypoints);
	constallFiles=getAllFiles('./build/aws/dist-serverless/nodejs/node_modules');
	constfilesToDelete=allFiles.filter(file=>!fileList.has(file));

Fix in Cursor • Fix in Web

---

Was this report helpful? Give feedback by reacting with 👍 or 👎