feat(trace-view): Add Server Rendering section to pageload span details by s1gr1d · Pull Request #130769 · getsentry/sentry

s1gr1d · 2026-11-11T10:39:27Z

A pageload of a static or partially static page can come from HTML that the server rendered earlier for other page loads. The server spans of that render are in another trace.

The new section reads page.server_rendering.type and the cache_origin and server_request span links. It shows which part of the page was reused and which part was rendered for this page load, and links to the trace of each.

The section is behind performance-trace-cache-lifecycle.

"Show Spans" will highlight the dynamic server span within the trace

When both parts are in another trace (e.g. the dynamic server part arrived too late):

Some trace view example links can be seen in the other PR in the stack: #130771

A pageload of a static or partially static page can come from HTML that the server rendered earlier for other page loads. The server spans of that render are in another trace, so users cannot find them from this trace. The new section reads page.server_rendering.type and the cache_origin and server_request span links. It shows which part of the page was reused and which part was rendered for this page load, and links to the trace of each. It works for every SDK that sets these attributes, whatever the framework. Move the linked span hook out of the cache lifecycle section, so both sections use it. The section is behind performance-trace-cache-lifecycle.

s1gr1d · 2026-11-11T10:40:25Z

bugbot run

cursor

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{❌ Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}

Want fixes drafted automatically? Bugbot Autofix can create code changes for findings. A team admin can enable Autofix in the Cursor dashboard.

^{Reviewed by Cursor Bugbot for commit 152250c. Configure here.}

…uest Go to span took the first http.server span below the pageload. That can be an API request of the page. Use the span that the tree moved below the pageload as the server request of the page.

nsdeschenes

Two small nits to take a peek at, rest LGTM 🚢

nsdeschenes · 2026-11-12T11:44:17Z

+  const cacheOrigin = links?.find(
+    link => getSpanLinkType(link) === CACHE_ORIGIN_LINK_TYPE
+  );
+  const serverRequest = links?.find(
+    link => getSpanLinkType(link) === SERVER_REQUEST_LINK_TYPE
+  );


nit: Can we pull out these calls to getSpanLinkType(link) out of these .find() calls, and assign to a variable then just pass that variable.

Something like:

const linkType = getSpanLinkType(link) const cacheOrigin = links?.find( link => linkType === CACHE_ORIGIN_LINK_TYPE ); const serverRequest = links?.find( link => linkType === SERVER_REQUEST_LINK_TYPE );

good point but your proposed helper does not work as link only exists in the callback. but I'll add a small helper.

Ahhh shoot sorry i misread it 🤦

nsdeschenes · 2026-11-12T11:44:17Z

+/** Dashed, because the spans of a reused part are never in this trace. */
+const ReusedBlock = styled(Stack)`
+  border: 1px dashed ${p => p.theme.tokens.border.primary};
+`;


nit: Looking at the docs, i think you should be able to pass this directly to the border property on the Stack. Though you may need to call useTheme to get the theme value 🤔

This is just part of an effort to maintain the reduction of styled calls throughout the app, if the suggestion doesn't work we can leave as is.

The border prop only renders a solid line but the dashed line is on purpose to give more meaning to "this span is not in this trace". But I can use border=primary and a style prop to override that.

chargome · 2026-11-12T12:47:40Z

+  });
+
+  return (
+    <Link to={linkedSpan.to} onClick={linkedSpan.onClick}>


Non-blocking: for pages rendered at runtime and reused for more than ~1.5 days (long/no revalidate), this link opens an empty trace because the query only searches around the pageload start right? Is it worth hiding the link or hinting at it in that case, or are we fine leaving it?

Yes, I'll add a follow-up where we can calculate the render time relibably.

the conventions PR: getsentry/sentry-conventions#799

Find the cache_origin and server_request links with one helper. Draw the reused block with the Stack border prop. The prop only draws solid lines, so the dashed style stays as an inline style.

… trace (#130771) Stacked on #130769 When the server reuses HTML for a page (a static page, or the static shell of a partially static page), the server spans of that render are in another trace. Until now, the waterfall showed nothing in their place, so it looked like instrumentation was missing. This adds a placeholder row under the document request of each pageload with `page.server_rendering.type`: - **No bar on the timeline.** The render can be hours older than the page load, so the row fills the timeline column with a pattern and the text "Rendered earlier: not on this timeline". - **Visible when the pageload is collapsed,** next to the server request, like a server transaction. For a fully static page: https://sentry-4x3dkn58u.sentry.dev/explore/traces/trace/cdf98425fa8746f79352a8394e385f32/?node=sr-85846057874639b8-server-render&timestamp=1794319075 <img width="1375" height="387" alt="image" src="https://github.com/user-attachments/assets/9023504d-1b83-4f7a-9329-4de0c1b6a766" /> For a partially static page: https://sentry-4x3dkn58u.sentry.dev/explore/traces/trace/d5f46a9d8bc640ca8d72c415f1b9b0a7/?fov=0.19530032969846994%2C71.30445552967653&node=span-9e805f433be01de5&timestamp=1794326348 <img width="1379" height="491" alt="image" src="https://github.com/user-attachments/assets/a63dddcb-0f7a-49b8-b1fc-54a73f492c82" />

s1gr1d requested a review from a team as a code owner November 11, 2026 10:39

github-actions Bot added the Scope: Frontend Automatically applied to PRs that change frontend components label Nov 11, 2026

cursor Bot reviewed Nov 11, 2026

View reviewed changes

Comment thread .../app/views/performance/traceDetails/traceDrawer/details/span/eapSections/serverRendering.tsx Resolved

s1gr1d added this pull request to stack #130772 November 11, 2026 10:54

s1gr1d mentioned this pull request Nov 11, 2026

feat(trace-view): Show where server spans are for reused pages in the trace #130771

Merged

s1gr1d requested review from chargome and nsdeschenes November 11, 2026 11:16

fix(trace-view): Go to the server request of the page, not an API req…

39b80c3

…uest Go to span took the first http.server span below the pageload. That can be an API request of the page. Use the span that the tree moved below the pageload as the server request of the page.

vercel Bot deployed to Preview November 11, 2026 14:05

View deployment

nsdeschenes approved these changes Nov 12, 2026

View reviewed changes

chargome approved these changes Nov 12, 2026

View reviewed changes

ref(trace-view): Address review comments on the Server Rendering section

20bbdc2

Find the cache_origin and server_request links with one helper. Draw the reused block with the Stack border prop. The prop only draws solid lines, so the dashed style stays as an inline style.

vercel Bot deployed to Preview November 13, 2026 09:49

View deployment

s1gr1d merged commit 3760ed5 into master Nov 13, 2026
83 checks passed

s1gr1d deleted the sig/ssr-trace-connection branch November 13, 2026 10:20

s1gr1d mentioned this pull request Nov 13, 2026

feat(attributes): Add page.server_rendering.age attribute getsentry/sentry-conventions#799

Open

4 of 5 tasks

Navigation Menu

Repository navigation

feat(trace-view): Add Server Rendering section to pageload span details - #130769

feat(trace-view): Add Server Rendering section to pageload span details#130769
s1gr1d merged 3 commits into
masterfrom
sig/ssr-trace-connection

s1gr1d commented Nov 11, 2026 •

edited

Loading

Uh oh!

s1gr1d commented Nov 11, 2026

Uh oh!

cursor Bot left a comment

Uh oh!

Uh oh!

nsdeschenes left a comment

Uh oh!

nsdeschenes Nov 12, 2026

Uh oh!

s1gr1d Nov 13, 2026

Uh oh!

nsdeschenes Nov 13, 2026

Uh oh!

nsdeschenes Nov 12, 2026

Uh oh!

s1gr1d Nov 13, 2026

Uh oh!

chargome Nov 12, 2026

Uh oh!

s1gr1d Nov 13, 2026

Uh oh!

s1gr1d Nov 13, 2026

Uh oh!

Uh oh!

This branch was successfully deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

Conversation

s1gr1d commented Nov 11, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

s1gr1d commented Nov 11, 2026

Uh oh!

cursor Bot left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

nsdeschenes left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

This branch was successfully deployed

Reviewers

Uh oh!

Assignees

Labels

Projects

Milestone

Development

Uh oh!

3 participants

s1gr1d commented Nov 11, 2026 •

edited

Loading