What is the difference between synthetic monitoring and real-device monitoring?
Synthetic monitoring runs scripted checks from servers in datacenters, usually with a headless browser. It is cheap and frequent, and it tells you whether your site answers a datacenter. Real-device monitoring loads the same page on a physical phone on a carrier network, so it also catches what only real mobile users meet: blocks by user agent or location, bot challenges, 403 pages and mobile-only redirects. Witnio's Access Monitor runs these checks on real Android phones and records DNS, TLS, HTTP result, redirects, challenge detection, load time and a signed screenshot.
Updated September 25, 2026
What synthetic monitoring does well
Datacenter checks are the right tool for broad availability coverage. They run every minute from many locations, cost little per check and give a consistent baseline for comparing releases. If your origin server goes down, a synthetic check will usually be the first to notice.
Their limit is the vantage point. A headless browser on a cloud IP address is exactly the kind of client that bot protection, firewalls and fraud rules are built to treat differently. The check can pass while real customers on phones get a challenge page, or fail while customers are fine.
What only a real device sees
A physical phone on a carrier network presents a mobile IP address, a real mobile browser and real device characteristics. That is the combination your mobile customers present, so the page it receives is the page they receive: the same CAPTCHA, the same geo block, the same broken mobile redirect.
The trade-off is cost and frequency. Real phones are finite, so real-device checks run less often than datacenter checks. The two methods complement each other: synthetic checks for fast, broad coverage, real-device checks for the critical pages where the mobile path is what matters, such as checkout, login and landing pages.
What a real-device check should record
A result is only useful if it explains itself. The minimum is DNS resolution, the TLS handshake, the HTTP status, the full redirect chain, whether a challenge or 403 page appeared, the time to load and a screenshot of what was on screen. With those fields a failure points to a layer, not just to a red dot.
It also matters who failed. If the site is down, a correctly classified and evidenced failure is a valid result. If the monitoring phone failed, the honest result is inconclusive, and it should not open an incident against the site or count against its availability.
How Witnio measures it
You provide the URL and a check frequency. A physical Android phone on a real SIM loads the page and records DNS, TLS, HTTP result, redirect chain, challenge or 403 detection, time to load and a screenshot. Each run is signed with Ed25519 and can be verified on the public verify page. Screenshots are kept 60 days on Founder and 90 days on Standard.
Access Monitor checks technical availability, not business content: a page that loads correctly but shows the wrong price is out of scope. System reliability was measured at 2,403 qualified backtests with 99.89% autonomy (2026-09-15), under injected faults such as network loss, reboots and service restarts.
Frequently asked questions
- Should real-device monitoring replace synthetic monitoring?
- No. Use synthetic checks for broad, frequent coverage and real-device checks for critical mobile paths where bot protection or location rules apply.
- Is Access Monitor run on an emulator?
- No. Every check runs on a physical Android phone with a real SIM, never an emulator and never a VPN exit.
- Does a screenshot replace functional testing?
- No. The screenshot shows what loaded; it does not test business flows such as completing a purchase.