Performance of NextNodeServer.createComponentTree span in OpenTelemetry traces

[DevWisdom8]

Summary

In my Next.js app, I’ve enabled instrumentation using OpenTelemetry, and I’m exporting traces to AWS X-Ray.

I noticed a span called NextNodeServer.createComponentTree that sometimes takes around 1 second to complete. The child spans are NextNodeServer.getLayoutOrPageModule, and there’s a large gap between them.

I’d like to understand:

What exactly happens during NextNodeServer.createComponentTree?

Are data fetching calls included in this span?

How can I optimize performance for this part of the trace?

Additional information

No response

Example

No response

[magictalent]

The NextNodeServer.createComponentTree span measures how the server prepares the React component tree before rendering.
It includes module resolution, layout loading, and preparing the server component boundaries, but not the data fetching itself — those happen after the tree is created.

If it takes ~1s, it’s often due to:
Heavy layout or middleware imports

Multiple dynamic imports or slow cold starts (especially on AWS Lambda)

Lack of persistent caching (cacheComponents: true helps)

To optimize:

Use use cache for static routes

Reduce imports in layout.tsx

Ensure your Lambda has warm concurrency or a smaller bundle size

You can confirm this by checking the spans right after the tree creation — if getLayoutOrPageModule shows delay, it’s module resolution overhead.