Performance

Response time, throughput, and Lighthouse-based performance metrics integrated into your test runs.

What this module tests

The Performance module measures and asserts on your application's performance characteristics. Rather than load testing, this module focuses on baseline performance validation — ensuring each release meets defined performance budgets. Tests cover:

• API response times — p50, p95, p99 latency per endpoint
• API throughput — requests per second at defined concurrency levels
• Web page performance metrics — Lighthouse scores (Performance, LCP, FID/INP, CLS, TTFB)
• Performance regression detection — alerting when metrics degrade beyond threshold vs. baseline

Metrics reference

Metric	Applies to	Default threshold
Response time p95	API	2000ms
Response time p99	API	5000ms
Throughput	API	No default — set per scenario
Lighthouse Performance score	Web	≥ 70
LCP (Largest Contentful Paint)	Web	≤ 2500ms
CLS (Cumulative Layout Shift)	Web	≤ 0.1
TTFB (Time to First Byte)	Web	≤ 800ms

Default thresholds can be overridden per project in Project Settings → Performance Thresholds or per scenario using max_response_ms and min_throughput_rps KB fields.

KB scenario format for performance tests

# Performance scenario examples
domain: products-api
scenarios:
  - id: PERF-API-001
    title: "GET /products responds within 500ms at p95"
    module: performance
    endpoint: /api/v1/products
    method: GET
    max_response_ms: 500
    steps:
      - Send GET to /api/v1/products with valid auth
    expected: >
      HTTP 200 within 500ms at p95 latency.
      p99 latency below 1000ms.

domain: homepage
scenarios:
  - id: PERF-WEB-001
    title: "Homepage Lighthouse performance score ≥ 80"
    module: performance
    steps:
      - Navigate to / (homepage)
      - Wait for page to be fully loaded
    expected: >
      Lighthouse Performance score ≥ 80.
      LCP ≤ 2500ms.
      CLS ≤ 0.1.

Performance baselines and regression alerts

After the first run, VigilQA records performance baselines. On subsequent runs, if a metric degrades beyond the configured regression threshold (default: 20% worse than baseline), a performance regression alert is raised. This alert appears in the run results and triggers a notification (if configured).