What is synthetic monitoring?
Synthetic monitoring runs scripted checks against your site or API on a schedule from outside, imitating what a user or client does: load a page, call an endpoint, log in, add to cart. Because the checks run whether or not anyone is using the service, they detect outages at night and test journeys before real users hit them.
Why synthetic monitoring matters
It is the most direct answer to "can users use it right now?". Simple checks confirm the site answers; multi-step API flows and browser scripts confirm that login, search or checkout still work after a deploy.
Synthetic results are consistent (same script, same place), which makes them good for SLAs and trends. Their limit is that they only test the paths you script, so pair them with real user monitoring.
How to measure synthetic monitoring
checks per month = locations × (minutes in the month ÷ interval in minutes) worst-case detection ≈ interval + confirmation delay
Worked example
An uptime check runs every minute from 3 locations, and a browser check of the checkout journey every 10 minutes from 1 location, over a 30-day month (43,200 minutes).
- Uptime checks: 3 × (43,200 ÷ 1) = 129,600 checks a month.
- Browser runs: 43,200 ÷ 10 = 4,320 runs a month.
- A full outage is noticed within about one minute plus the confirmation retry; a broken checkout only within up to 10 minutes, the browser check's interval.
Result: Spend frequent, cheap checks on availability and fewer, richer browser runs on the journeys that make money.
Common mistakes
- Only checking the homepage. The homepage is often cached and the last thing to break.
- Brittle scripts. Selectors tied to layout break on every redesign. Use stable IDs or visible text.
- Test accounts with real side effects. A synthetic checkout must not create real orders or charge cards. Use test modes and clean up.
- Alerting on one failure. Confirm before alerting, or network blips page people.
How Vigil handles synthetic monitoring
- Vigil covers the range: Uptime checks (with SSL expiry), API checks with JSON, text and header assertions, multi-step API flows that pass tokens between steps, TCP port checks, and Browser checks that run a short script in headless Chromium with screenshots as evidence.
- Failures are confirmed by a second check about 12 seconds later, and private locations run checks from inside your own network.
- Try a one-off check with the free uptime checker or the HTTP request tester.
Synthetic monitoring: frequently asked questions
What is the difference between synthetic monitoring and RUM?
Synthetic monitoring runs scripted checks on a schedule; RUM measures real visitors. Synthetic detects outages with no traffic and gives consistent baselines; RUM shows the real spread of devices and networks.
How often should synthetic checks run?
Every 1 to 5 minutes for critical endpoints, less often for expensive browser journeys. The interval is your worst-case detection time.
Is an uptime check a synthetic check?
Yes, the simplest kind: one request and a check of the response. API flows and browser scripts are the more advanced kinds.