This introduces a new experimental Sentry Lambda extension within the existing Sentry Lambda layer. Sentry events are being tunnelled through the extension, where they are then forwarded to Sentry.

Initial benchmarks using ApacheBench show a reduction in request processing time of around 26% (from 243ms to 180ms on average over 100 requests; function pre-warmed).

To enable it, set _experiments.enableLambdaExtension in your Sentry config like this:

Sentry.init({
 // ...other config
 dsn: "<YOUR_DSN>",
 _experiments: {
 enableLambdaExtension: true
 }
})

closes #12856
relates to #3051