Standard RUM (Datadog, New Relic) relies on a healthy connection to load its monitoring script. EdgeViz turns every browser into a persistent sentinel that maximizes path diversity, evolving new ways to get diagnostic signals through even when the primary HTTP path is completely broken.
Traditional monitoring has two fatal flaws. Synthetics check from a handful of clean datacenter backbones, missing regional "last-mile" ISP failures. Standard RUM relies on a JavaScript snippet that never loads if your site is down. EdgeViz uses a persistent Service Worker and out-of-band signaling to capture the ground truth from actual user browsers.
| capability | edgeviz | standard rum | legacy synthetic |
|---|---|---|---|
| works when site is 100% down | ✓ | ✗ | ✓ (if origin is up) |
| out-of-band signal delivery | ✓ | ✗ | ✗ |
| real user ISP/ASN breakdown | ✓ | ✓ | ✗ |
| edge PoP diagnostics | ✓ | ✗ | ✗ |
| offline UI injection | ✓ | ✗ | ✗ |
| same-origin hosting | ✓ | ✗ | N/A |
Legacy monitoring depends on cookies and behavioral data collection. EdgeViz captures network diagnostics only — no cookies, no query strings, zero external dependencies — reducing your attack surface and compliance burden.
Service worker catches connection failures. Detects origin timeouts, DNS failures, TLS errors and 5xx error pages—from the user's actual network path. Only the failed page's origin and path are recorded.
Every beacon includes ISP, ASN, edge colo, TCP RTT, TLS version. Relay observations add JA4 fingerprints and independent geo cross-checks. Pinpoint the exact network path where failures occur.
Beacons grouped by domain + failure type. In a major outage, one alert names every open incident instead of a flood of them.
Auto-classifies: origin down, origin error, client connectivity, partial failure. Edge worker probes origin to confirm server-side vs client-side.
When affected sessions reconnect, incidents update. See real-time recovery rates across networks.
Visitors get a status page with live diagnostics instead of the browser's error, and it reloads itself when the site is back.
Multiple independent signal paths across diverse infrastructure providers, operating independently of your primary HTTP/TLS stack. If the primary route is blocked, diagnostics route through secondary channels. No single point of failure between the user and your visibility.
Cross-references client location with edge PoP distance across multiple independent relay providers. Surface bad routing, wrong datacenter hits, and unreliable geo data with cross-provider verification.
No cookies. No query strings. No page content. We capture network diagnostics only—ISP, edge, timing—and the failed page's origin and path. Your users' data stays yours.
Service workers require same-origin hosting. Add our proxy rules, then embed the script.
location ~ ^/(edgeviz-sw|edgeviz-core|edgeviz)\.js$ {
proxy_pass https://edgeviz.net;
proxy_ssl_server_name on;
}<script async src="https://yourdomain.com/edgeviz.js?id=YOUR_ID"></script>Goes on every page — a visitor's first hit is as likely to be a deep link as the home page. Full docs with Apache, Caddy, Cloudflare, and Vercel configs.
Registers service worker (~38KB). No external dependencies.
Watches for connection failures and 5xx error pages only. A healthy page is handed over as it arrives. No data inspection.
Request times out or errors. SW captures network context via edge worker probe.
Diagnostic payload finds a path to edgeviz—even during partial outages. Aggregated into active incident.
Live push to your dashboard. Incident cards populate with network forensics as beacons arrive.
When 500 users across different sites report failures from Verizon in Chicago at the same time, that's not your origin—it's an ISP outage. EdgeViz correlates telemetry across all customers to detect infrastructure failures before ISPs admit them.
Don't take our word for it. See exactly how EdgeViz intercepts a connection failure and provides real-time forensics in our interactive simulation.
No credit card. One line of code. Full visibility.