You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
MSTest.Performance.Runner is a capable local profiling harness (PerfView, dotnet-trace, timing via PlainProcess), but it has no CI integration. Performance regressions can be merged silently because:
No automated timing baseline is checked on every PR or nightly.
There is no trend tracking (no history of "how long does Scenario1 take to run?").
Proposal
Add a lightweight CI step — ideally on a Linux runner since it's the most reproducible — that:
Runs ./build.sh -pack to produce the MSTest NuGet packages.
Uploads Result.json as a CI artifact for trend tracking.
(Optional, phase 2) Compares the timing against a stored baseline and fails the check if regression exceeds a threshold (e.g. >10% slowdown).
Phase 1 (artifact-only) is low-risk: it does not gate PRs and only adds visibility. Phase 2 (regression gate) should be discussed with maintainers before implementation.
Why now
PR #9311 removes the Windows-only restriction from the timing pipelines, making a Linux CI runner a viable host for the first time. This issue tracks the follow-up work to wire it into the CI pipeline.
Effort estimate
Phase 1: small — a new GitHub Actions job, ~1 day of work.
Phase 2: medium — requires agreed baseline values and acceptable-regression thresholds.
🤖 Automated content by GitHub Copilot. Posted via a maintainer's GitHub token, so it appears under their account — the account owner did not write or approve this content personally. Generated by the Perf Improver workflow. · 1.3K AIC · ⌖ 25.8 AIC · ⊞ 57.6K · [◷]( · ◷)
Add this agentic workflows to your repo
To install this agentic workflow, run
gh aw add githubnext/agentics/workflows/perf-improver.md@main
Activity
Add Phase 1 nightly perf-timing CI workflow (#9312) (#9325)