Synthetic Monitoring
Lighthouse and TTFB lab tests on your pages, side by side with real-user field data.
Synthetic Monitoring measures your pages in a lab (scheduled Lighthouse runs plus lightweight TTFB checks) and shows the results next to your real-user (RUM) field data. The point isn't the lab number by itself; it's spotting where lab and field disagree, and knowing which one to trust.
Beta
Synthetic Monitoring is in beta. Any member of the organization can use it, but response shapes and UI may still change.
Lab vs. field, in one sentence each
- Lab (synthetic): one simulated device, one location, ideal conditions, repeatable. Great for catching regressions and debugging, but blind to what real visitors experience.
- Field (RUM): every real visitor, real networks, real devices, aggregated at p75. The truth, but noisy and slower to move.
Fastmon shows both on the same charts so neither gets mistaken for the other.
Turning it on
Synthetic Monitoring is off by default and is enabled per organization: Organization Settings → Synthetic Monitoring → External measurements. Only the organization owner can flip it.
It's opt-in because the measurements run on an external provider (Commerce-Score) that acts as a sub-processor: the URLs you monitor, and only those, are sent there. If your DPA review needs the details, that's the one data flow to look at. Visitor data never leaves fastmon; only the page URLs you explicitly add do.
Pages
A synthetic page is one URL under your site's domain that gets measured. Three ways a page comes to exist:
| Source | How | Counts against |
|---|---|---|
| Manual | You add the URL yourself | Manual page quota |
| Auto | "Auto-fill" picks your top URLs by RUM traffic (trailing 28 days, up to 5 new pages per click) | Auto page quota |
| One-shot | A single ad-hoc test, no recurring schedule | Nothing |
Quotas are per organization (across all its sites), with separate budgets for manual and auto pages:
| Plan | Manual pages | Auto pages |
|---|---|---|
| Beta | 5 | 5 |
| Starter | 2 | 3 |
| Pro | 5 | 10 |
| Partner | 10 | 20 |
| Enterprise | 25 | 50 |
At the cap, adding a page fails with a clear error; delete a page to make room. One-shots are deliberately unlimited; use them to answer "how does this page score right now?" without spending budget. A one-shot can be promoted to a monitored page later.
Schedules and per-page settings
Every page has three switches, in the page's settings dialog:
- Recurring Lighthouse run: a full Lighthouse measurement every 24 hours. Turn it off to pause; the page stays and can still be run on demand.
- Recurring TTFB run: a lightweight server-response check roughly every 5 minutes, independent of the Lighthouse schedule.
- Bypass cache (default on): busts the measured page's CDN and browser cache on each run, so you measure the worst-case, uncached load. Turn it off to measure the cached path instead.
On top of the schedules, Send synthetic test triggers an immediate run for any page.
What a run gives you
Every run measures desktop and mobile in one go.
- Lighthouse report: category scores (Performance, Accessibility, Best Practices, SEO), the lab metrics (LCP, FCP, CLS, TBT, Speed Index, TTFB), opportunities sorted by estimated savings, request waterfall, third-party and resource breakdowns, and a lab screenshot per device. The raw Lighthouse JSON is one click away if you want everything.
- Lab vs. field: lab LCP/FCP/CLS/TTFB against your RUM p75 for the same URL, each with a verdict: aligned, field slower, or field faster. Field INP is shown without a lab twin (Lighthouse can't measure INP; lab TBT is the closest proxy). Verdicts need at least 50 field samples in the window; below that they're marked insufficient rather than misleading.
- Experience Score vs. lab score: your field Experience Score next to the lab composite score, same caveat: they measure different things; disagreement is signal, not error.
- TTFB diagnosis: the dedicated TTFB runs break server response into phases (DNS, TCP, SSL, server, transfer) and compare against field TTFB, with a verdict on whether you're server-bound or network/geo-bound.
- Trends: lab points (one per run) overlaid on the regular field p75 line, plus the 5-minute TTFB series, over a selectable window.
AI summaries
Two AI features sit on top of the Lighthouse data, both gated by the organization's AI setting:
- One-click summary: turns a Lighthouse report into a short written summary, in English or German. Generated once, stored, and reused; generation is rate-limited to 30 per organization per hour (reading a stored summary is free).
- Ask about this test: opens the fastmon AI assistant with the run in context, so you can ask follow-up questions about specific audits.
API
Everything above is available programmatically under /v1/sites/{site_id}/synthetic/…: pages CRUD, on-demand runs, results, timeseries, lab-vs-field, screenshots, and the AI summaries. See the Synthetic endpoint reference.
What's next
- Experience Score: the field composite that lab runs are compared against.
- Web Vitals overview: what LCP, INP, CLS actually measure.
- Analytics: the field-data side of the same story.