fix: Running in Lambda has inherent 6mb request size limit

[yhay81]

Fixes #677.

Summary

Running in Lambda has inherent 6mb request size limit

Validation

• Mechanical gate: +6/-0, quality and tests passed
• Adversarial review: approved

---

🤖 AI-authored PR, operated by @yhay81.

[pullfrog]

ℹ️ Minor suggestions only.

Reviewed changes

• Lambda 6 MB payload-limit warning: adds a > **Warning:** blockquote to the top of the Lambda deployment guide documenting the 6 MB synchronous invocation payload limit, that artifacts larger than 6 MB cannot be uploaded, and that Turbo reports a 413 Payload Too Large warning. (+6 lines, docs-only, resolves #677.)

The core facts check out: the 6 MB synchronous request/response quota and the cited AWS docs page are correct, and 413 Payload Too Large matches the error the issue reporter actually observed from a Function URL. One inline suggestion on the effective artifact-size boundary.

  ^{｜ Fix all ➔ ｜ Fix 👍s ➔ ｜ View workflow run ｜ Using DeepSeek Flash (default — pick a model for stronger reviews) ｜ 𝕏}

[matteovivona]

Thanks for picking this up. One correction needed before merge, plus three things worth adding while we are here.

1. src/aws-lambda.ts:8 sets enforceBase64: (_) => true, and Lambda Function URLs base64-encode binary request bodies (Turbo uploads with application/octet-stream). Base64 adds about 33%, so a 6 MB invocation payload carries only ~4.5 MB of artifact bytes.

2. The sentence names the response limit but then says only that artifacts "cannot be uploaded". With enforceBase64 on the response path, a GET of an oversized artifact fails the same way. Worth stating both directions.

3. Anyone who hits this will reach for BODY_LIMIT (docs/environment-variables.md:28, default 100 MB). It does nothing here. AWS rejects the request before Fastify sees it. One sentence saves that detour.

4)TURBO_CACHE_READ_URL (docs/environment-variables.md:34) answers reads with a 302 to a CDN instead of streaming through the server, which removes the response-side limit entirely. Writes still hit the 6 MB cap, but it is a genuine partial fix this repo already supports.

[matteovivona]

@all-contributors add @yhay81 on docs

[allcontributors]

@matteovivona

I've put up a pull request to add @yhay81! 🎉

[matteovivona]

🎉 This PR is included in version 2.12.2 🎉

The release is available on:

• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀