feat(ci): automated performance regression detection in CI pipeline

[thelmaoffiong]

Automated Performance Regression Detection in CI Pipeline

Closes #171

Problem

Performance regressions can silently slip into the frontend — a heavier bundle, a slower API handler, or an expensive render path degrades P99 latency of critical paths without breaking any functional test. This PR adds an automated performance regression detection pipeline that runs in CI and blocks merges when critical paths regress.

What's included

1. Solution architecture & design

• docs/performance-regression-detection.md — architecture doc covering the measurement → analysis → gate flow, budgets, baseline lifecycle, monitoring/alerting/dashboards, and blue-green/canary analysis support.

2. Core logic with comprehensive tests

• src/utils/performanceRegression.ts — pure, deterministic detection logic: nearest-rank percentiles, budget evaluation, and baseline-relative regression detection with configurable tolerance.
• tests/unit/performanceRegression.test.ts — 24 unit tests (100% statement/function/line coverage, 92% branch coverage).

3. Monitoring, alerting, and dashboards

• tests/performance/performance.spec.ts — Playwright benchmark measuring API critical paths (/api/runtime-config/audit, /api/rate-limit) with a P99 < 100ms budget plus page TTFB/LCP for / and /export.
• scripts/analyze-performance.ts — summarizes samples, evaluates budgets + baseline, emits an HTML dashboard (report.html), a machine-readable report (report.json), and a PR comment; exits non-zero to block the CI gate.
• .github/workflows/performance-regression.yml — runs on PRs (posts findings comment + uploads artifacts) and refreshes the committed baseline on main.

4. Blue-green strategy and canary analysis

• The same baseline-comparison mechanism powers canary analysis: green (candidate) measurements are compared against the blue baseline before promotion, mirroring the thresholds in docs/TRACING_DEPLOYMENT.md.

5. Runbooks and documentation

• docs/runbooks/performance-regression.md — triage, remediation, budget tuning, and escalation.

Gate behavior

Status	Meaning	Blocks merge?
pass	Within budget and baseline tolerance	No
budget-breach	Hard budget exceeded (e.g. P99 > 100ms)	Yes
regression	Slower than committed baseline by > tolerance	Yes
insufficient-data	Too few samples	No

Verification

• npx tsc --noEmit ✅
• npm run lint ✅
• npx vitest --run — 678 tests passing (84 files) ✅
• End-to-end analyzer verified locally with sample data for pass / budget-breach / regression paths ✅