Pagespeed test (server-side)
3-run measurement of connect, TLS, TTFB and total. Shows HTTP version, gzip/brotli and body size — straight from the XICTRAQ node.
How this tool works
The pagespeed test combines server RTT measurement (3 runs for TCP connect, TLS handshake, TTFB, total) with real web vitals from a headless Chromium run: LCP, CLS, FCP and TBT graded against web.dev/vitals. Three Lighthouse runs produce a median to filter single-run variance. The test also detects HTTP/2 vs HTTP/3 and gzip/brotli. You get both backend response time and rendering user experience in a single tool.
Typical use cases
- Compare hosting before/after — TTFB is the most honest metric
- Measure backend optimisations (caching, DB index, PHP OPcache) via TTFB
- Test CDN edges across regions
- Verify HTTP/2 / HTTP/3 enablement
- Check brotli compression after server config changes
Frequently asked questions
How does the test differ from Lighthouse?
We do both: server RTT (TCP, TLS, TTFB, total) over 3 runs and web vitals (LCP, CLS, FCP, TBT) from a headless Chromium run with median over 3 iterations — same engine as Lighthouse, just without the full Lighthouse report. For the complete Lighthouse report including SEO/Accessibility/Best-Practices score, use the XICTRAQ compliance scan in the dashboard.
Which TTFB is good?
This is a server response measured from a data centre, not the value a real visitor experiences - the thresholds are correspondingly tight: under 200ms excellent, 200-400ms acceptable, above 800ms notable. Google's Core Web Vitals thresholds apply to field data from real visitors and are wider: up to 800ms good, 800-1800ms needs improvement, above that poor. Lighthouse flags a server response time from 600ms as worth addressing.
Why three runs instead of one?
The first run cold-starts TCP/TLS caches. Three runs show the variance — with warm cache the value should ideally drop to half of the first.
From which location is the measurement?
From our server in a data centre in Nuremberg (Germany). For global measurements, use the multi-node option in our monitoring (two nodes in Germany, one in Finland).
Does the test support HTTP/3 (QUIC)?
We detect HTTP/3 responses via Alt-Svc but currently measure only over HTTP/2. Full QUIC measurement is on the roadmap.
Related tools
Continuous monitoring vs. one-shot?
XICTRAQ monitors uptime, SSL, security headers, GDPR and performance 24/7. Free plan, no credit card.