← Blog

How to check page speed regularly (and why it is not uptime)

One run on pagespeed.web.dev after deploy helps, but it will not catch a regression next week: a heavier script in production, a CDN change, or bloated hero images on marketing pages. For public URLs (home, pricing, blog) you want repeatable measurements with history — separate from “does the server return 200?”

StillOnline still answers availability (HTTP health checks, status page, owner alerts). The project PageSpeed tab is about load speed and Core Web Vitals for public pages. Below: a practical monitoring rhythm and how our UI compares to other tools.

Quick answer

StillOnline PageSpeed (Pro/Ultimate) schedules Google PSI runs on public marketing URLs—mobile and desktop history in the same project as uptime checks. It does not replace HTTP /health monitoring or appear on your public status page. Point Page Speed at /, pricing, or blog—not /app or API health. Uptime alerts still follow checks (Telegram); there is no “Performance < 50” push by design.

What to measure on a schedule

Page Speed in StillOnline is for marketing surfaces your customers load without logging in—not for API JSON or authenticated app shells. If a URL needs a session cookie, keep it on the Monitoring tab instead.

ScenarioWhy repeat
Marketing / pricing after a releaseCompare mobile vs desktop before and after
Blog or docs on a CDNCatch LCP regressions from images or fonts
Storefront without loginURL must work without signing in — otherwise PSI does not reflect real visitors

Do not point Page Speed at /app/dashboard or your API health JSON. Page Speed measures page render; keep /health on the Monitoring tab — see health check quickstart.

Lab (Lighthouse) vs field (CrUX)

Google PageSpeed Insights returns two layers:

  • Lab — synthetic Lighthouse: Performance, FCP, LCP, CLS, TTFB. Throttling is consistent → good for charts and post-deploy comparison.
  • FieldINP from CrUX: p75 of real Chrome visits over ~28 days. Low URL traffic may fall back to origin-level data; “—” means INP was not returned.

Search Console shows field CWV with lag and without lab history next to your uptime project. A one-off PSI tab has no account history.

The PageSpeed tab in StillOnline

Open PageSpeed in the project (next to Overview, Monitoring, and History).

Snapshot on Overview

Overview shows the latest Performance scores (mobile / desktop) and a small sparkline. PageSpeed → jumps to the full tab. On Free you see an upgrade hint; full runs require Pro or Ultimate.

Full PageSpeed tab (Pro / Ultimate)

  1. Metrics guide — lab vs CrUX, links to web.dev.
  2. Chart — pick page and metric (Performance, FCP, LCP, INP, CLS) over time.
  3. Add page — label + public HTTPS URL (3 URLs on Pro, 5 on Ultimate).
  4. URL cardMobile and Desktop panels: score, FCP, LCP, INP, CLS, TTFB; Run button; link to the full PSI report.
  5. Recent runs table — time, device, auto vs manual.

Automatic runs: mobile + desktop every 24h (Pro) or 12h (Ultimate). Manual runs: 1 or 3 per URL per day (UTC day). Queue order: mobile then desktop; Google API rate limits are retried in the background.

Important:

  • Data source: Google PageSpeed Insights API (same as pagespeed.web.dev).
  • Not shown on your public status page and not a substitute for HTTP checks.
  • If project monitoring is paused (all uptime checks disabled), auto and manual Page Speed runs stop until you resume.
FreeProUltimate
PageSpeed tabpreview + upsellfull UIfull UI
URLs per project035
Auto mobile + desktop24h12h
Manual runs / URL / day13

Pro is $9/mo with uptime, status pages, and API — pricing. Start free for uptime; add Page Speed after upgrade.

Quick setup

  1. Project with uptime monitoring active (not paused).
  2. PageSpeedAdd page → e.g. https://yourproduct.com/ labeled “Home”.
  3. Run after deploy or wait for the next automatic run.
  4. Compare mobile vs desktop; open Full PageSpeed Insights report for Lighthouse details.
  5. Watch the LCP/CLS chart for regressions before support tickets.

Owner alerts (email, Telegram, Slack) follow uptime checks — Telegram guide. Page Speed does not send separate “score dropped” pushes by design (noise vs “site is down”).

Comparison with other services

StillOnline does not replace SpeedCurve or full RUM platforms. We fit when you already run project + status page + health and want PSI with history in the same dashboard.

ToolStrengthStillOnline PageSpeed
pagespeed.web.devFree one-off reportSame API, plus schedule, history, chart, multiple URLs per project
Lighthouse CIPR/CI gatesNot in the pipeline — production URLs on a schedule next to uptime
GTmetrix / WebPageTestWaterfalls, regionsNo waterfall; CWV + uptime in one indie-friendly cabinet
SpeedCurve / CalibreRUM, budgets, perf teamsBroader and pricier; we are a light layer for 1–5 key URLs
UptimeRobot / Better StackMany pings, on-call, logsUptime-first; browser synthetic is a different product lane, not bundled PSI charts at $9
Search Console (CWV)Field data, SEO contextLab + INP (CrUX) with mobile/desktop after ship, without waiting 28 days for lab

Common stack: UptimeRobot or Kuma for volume, StillOnline for the status page + PageSpeed on home and pricing. Or StillOnline only if 10 uptime URLs on Pro and 3–5 speed URLs are enough.

Uptime tool pick: 2026 shortlist · comparison hub.

Related guides

FAQ

Does StillOnline Page Speed replace HTTP uptime monitoring?

No. StillOnline uptime checks ask “does this URL return success?” Page Speed asks “how fast does this public page render?” A slow 200 landing can stay green on uptime and poor on LCP—use both tabs. Keep /health on Monitoring; see health check quickstart.

Can StillOnline Page Speed monitor a URL behind login?

No. The URL must load without signing in, like your marketing home or pricing page. Dashboard and API routes belong on uptime Monitoring, not Page Speed—see public status page guide for what customers see externally.

Why does StillOnline not send alerts when Performance < 50?

StillOnline Page Speed focuses on history and post-release comparison, not threshold pushes (noise vs “site is down”). Uptime owner alerts on down still work via Telegram or Slack. Dedicated perf platforms handle budget alerting.

Can StillOnline Free users open the PageSpeed tab?

Yes—you see preview and upgrade copy on Free. Measured runs, charts, and scheduled PSI need Pro or UltimatePricing (3 URLs on Pro, 5 on Ultimate per project).

Does pausing StillOnline uptime monitoring stop Page Speed runs?

Yes. While all uptime checks in the project are paused, automatic and manual Page Speed runs stop until you resume monitoring in the StillOnline dashboard.