Skip to content

feat(app): show materialized view acceleration on the trace search RED charts - #3739

Draft
alex-fedotyev wants to merge 1 commit into
alex/HDX-6120-trace-rollup-3-exact-mv-edgesfrom
alex/HDX-6120-trace-rollup-4-mv-indicator
Draft

alex-fedotyev wants to merge 1 commit into
alex/HDX-6120-trace-rollup-3-exact-mv-edgesfrom
alex/HDX-6120-trace-rollup-4-mv-indicator

Conversation

@alex-fedotyev

@alex-fedotyev alex-fedotyev commented Nov 20, 2026 •

Copy link
Copy Markdown
Contributor

Stack

Layer 4 of 5 in a GitHub stack (stack #3741). GitHub shows each layer's base as the branch below it, but the stack's base is main: merging any PR lands it and every PR below it on main, one squashed commit each. Each PR's diff is only its own layer, and CI runs as if each targets main:

  1. fix(app): match the demo traces rollup's dimension columns to its table #3736 Demo traces rollup lists its real dimension columns
  2. fix(common-utils): merge DDSketch and GK quantile states with their accuracy #3737 DDSketch and GK states merge with their accuracy
  3. fix(app): keep search counts exact when a materialized view serves the chart #3738 Search counts stay exact when a view serves them
  4. feat(app): show materialized view acceleration on the trace search RED charts #3739 Acceleration indicator on the trace search RED charts (this PR)
  5. feat(app): infer a materialized view's minimum date from when it was created #3740 Minimum date inferred from the view's creation time

The layers are independent fixes to how charts read materialized views, ordered so the smallest land first. They prepare for a default trace rollup, which comes in a follow-up stack.

Summary

The Throughput, Errors and Duration tiles on trace search already read from a source's materialized views when one fits, but they hid the acceleration indicator that dashboard tiles show, so there was no way to tell whether a tile was accelerated or why not. They now show it, with the same dialog that lists which view served the chart or why each view couldn't.

The date-range indicator stays off, because the search range stays exact when a view serves the chart (see the PR below this one). The search histogram keeps its current look: it has no title row, so the indicator there would add a header to every search, including sources without views.

Screenshots

Last hour on the demo traces, where a 1-minute view serves all three tiles. Light theme:

RED tiles accelerated, light

Dark theme:

RED tiles accelerated, dark

The dialog from the Duration tile, naming the view that served it and why the other one couldn't:

Acceleration dialog

The default 15-minute range uses 15-second buckets, finer than the views, so the tiles read the source table and say so:

RED tiles not accelerated over 15 minutes

Test plan

  • Unit test: every RED tile renders with the acceleration indicator and without the date-range indicator.
  • e2e (materialized-views.spec.ts): on the E2E Traces MV fixture source over the last hour, all three tiles show Accelerated, and the dialog lists the view.
  • Checked on this stack's Vercel preview against the demo traces: over the last hour all three tiles report Accelerated in light and dark, and over 15 minutes all three report Not accelerated. The results count matched a raw count() over the same range exactly (3,077,342 and 3,078,092 for two one-hour windows, 762,061 for 15 minutes).
  • Viewport: the tiles keep the layout from the RED metrics PR; on a compact viewport the search page still shows the histogram, which has no indicator.
  • States: the indicator shows nothing for a source without views, Accelerated or Not accelerated otherwise. Loading and error states are the tiles' existing ones.

Part of HDX-6120.

@changeset-bot

changeset-bot Bot commented Nov 20, 2026 •

Copy link
Copy Markdown

🦋 Changeset detected

Latest commit: b24f1a0

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 3 packages
Name Type
@hyperdx/app Patch
@hyperdx/api Patch
@hyperdx/otel-collector Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

@vercel

vercel Bot commented Nov 20, 2026 •

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

Project Deployment Actions Updated
hyperdx-oss Ready Ready Preview Nov 20, 2026 6:31am UTC
hyperdx-storybook Ready Ready Preview Nov 20, 2026 6:31am UTC

Request Review

@github-actions

Copy link
Copy Markdown
Contributor

Duplicate code analysis

No new duplicate code introduced by this PR.


jscpd compared this branch against origin/main
across app, api and common-utils. Only new clones are listed; the codebase
currently holds 108 in total (1.13%).
Run yarn dupes locally to reproduce, or yarn dupes:all for the full picture.

@github-actions

github-actions Bot commented Nov 20, 2026 •

Copy link
Copy Markdown
Contributor

E2E Test Results

✅ All tests passed • 472 passed • 1 skipped • 1047s

Status Count
✅ Passed 472
❌ Failed 0
⚠️ Flaky 11
⏭️ Skipped 1

Tests ran across 4 shards in parallel.

View full report →

…D charts

The Throughput, Errors and Duration tiles already read from a source's
materialized views when one fits, but they hid the indicator dashboard tiles
show, so there was no way to tell whether a tile was accelerated or why not.
The tiles now show it; the date-range indicator stays off because the search
range is exact.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>

This branch was successfully deployed

2 active deployments
Preview – hyperdx-oss — b24f1a03 Deployed Nov 20, 2026 by vercel[bot]
Preview – hyperdx-storybook — b24f1a03 Deployed Nov 20, 2026 by vercel[bot]
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant