Skip to content

Conversation

Copy link
Contributor

Copilot AI commented Oct 28, 2025

Benchmark results vary naturally up to ~125%, causing false positives with the current 115% threshold.

Changes

  • CI workflow (.github/workflows/ci.yml): Updated alert-threshold from 115% to 130%
  • Documentation: Updated threshold references from outdated 150% value to 130% in:
    • benchmarks/README.md
    • benchmarks/PERFORMANCE_BASELINES.md (including 2.5x → 2.3x slowdown calculation)

This provides a reasonable buffer above natural variance while maintaining meaningful regression detection.

Original prompt

Change benchmark threshold to 130% because results often vary up to ~125%.


💬 We'd love your input! Share your thoughts on Copilot coding agent in our 2 minute survey.

Copilot AI self-assigned this Oct 28, 2025
Co-authored-by: LayZeeDK <6364586+LayZeeDK@users.noreply.github.com>
Copilot AI changed the title [WIP] Update benchmark threshold to 130% based on result variability Update benchmark regression threshold to 130% Oct 28, 2025
Copilot AI requested a review from LayZeeDK October 28, 2025 07:57
Copilot finished work on behalf of LayZeeDK October 28, 2025 07:57
@LayZeeDK LayZeeDK marked this pull request as ready for review October 28, 2025 08:04
@LayZeeDK LayZeeDK changed the title Update benchmark regression threshold to 130% ci: update benchmark regression threshold to 130% Oct 28, 2025
@LayZeeDK LayZeeDK enabled auto-merge October 28, 2025 08:05
@LayZeeDK LayZeeDK merged commit 8183942 into main Oct 28, 2025
14 checks passed
@LayZeeDK LayZeeDK deleted the copilot/change-benchmark-threshold branch October 28, 2025 08:11
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.

2 participants