fix(perf): show multi-region benchmark history - #901
Merged
Merged
Conversation
|
The latest updates on your projects. Learn more about Vercel for GitHub.
|
joshieDo
marked this pull request as ready for review
September 24, 2026 10:10
joshieDo
changed the title
The settled-TPS card now describes the latest multi-region benchmark, so its fallback should not show the old local transfer-only number. Use the Sep 29 multi-region public-mix nightly (~15.6K TPS). Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
The public stats now come from the multi-region public-mix nightly. Update the static copy that still described the old local transfer-only benchmark: 15k+ TPS headline, average execution fallback, a peak fallback no lower than the average, and a reliability caption without the old settled-load share. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Co-authored-by: decofe <256792747+decofe@users.noreply.github.com>
emmajam
approved these changes
Oct 1, 2026
Multi-region benchmarks average ~690ms block time, so stop promising settlement in half a second on the performance page. Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
emmajam
approved these changes
Oct 1, 2026
Thegreatsura
pushed a commit
to Thegreatsura/tempo
that referenced
this pull request
Oct 2, 2026
- Use shared `default` → `public-mix` (80/5/15; four tokens) for local, manual, and scheduled benchmarks. Record the concrete preset; local bloat defaults to 1024 MiB and rejects stale cached state. - Merge after [runner tempoxyz#432](tempoxyz/tempo-multi-region-benchmark#432) and [tempoxyz#434](tempoxyz/tempo-multi-region-benchmark#434), once [perf tempoxyz#178](tempoxyz/tempo-apps-internal#178) and [docs tempoxyz#901](tempoxyz/docs#901) are deployed. - Checked: preset and bloat-marker checks; earlier [e2e](https://github.com/tempoxyz/tempo/actions/runs/35889647294) and [AWS](https://github.com/tempoxyz/tempo-multi-region-benchmark/actions/runs/35889266582) smokes. --------- Co-authored-by: joshieDo <93316087+joshieDo@users.noreply.github.com> Co-authored-by: Alexey Shekhirin <5773434+shekhirin@users.noreply.github.com> Co-authored-by: Claude Opus 5.5 <noreply@anthropic.com>
This branch was successfully deployed
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
3 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Plot the scheduled multi-region benchmark history across presets as one time-scaled line, without settings/workload-change markers. Show history from Sep 1 on the performance page; drop the half-second settlement claim; keep homepage sparklines on the latest workload. Deploy after Perf #178, before Tempo #7848.
Checked: focused tests, types, build, and a local render against the live multi-region data served by the #178 preview. The Vercel preview shows fallbacks until #178 is deployed, because it reads production perf.tempo.xyz.
Live data (Oct 1 nightly), production vs. this branch:
Preview