Test your app from anywhere your users are.

Monitoring from outside your own network

Synthetic checks are tiny and frequent, so the cost driver is the number of vantage points rather than traffic:

Split the job. Whatever your target whitelists has to come from static addresses that never change, and that is a fixed monthly cost. Everything that tests what a user in a given country experiences belongs on residential, where two dollars buys a month of it.

Static where you are whitelisted, residential where geography matters

Mixing these two up is the usual failure. Whitelists need addresses that hold still, geography needs addresses that move:

Python: two paths, one monitor

STATIC = "http://login:password@198.51.100.20:50100"

def geo(country):
    return f"http://login_c_{country}:password@proxy.sotaproxy.com:10000"

def check(endpoint):
    if endpoint.whitelisted:
        return probe(endpoint, proxy=STATIC, timeout=10)
    return [probe(endpoint, proxy=geo(c), timeout=30) for c in endpoint.markets]

• Give whitelisted checks a shorter timeout than geography checks. A static route is fast and a slow answer is a real signal, while a residential route is slower by nature.
• Alert on a pattern, not on one failed probe. A single residential hop can fail for reasons that have nothing to do with your service.
• Keep at least two vantage countries for anything customer-facing. One region going dark while others stay green is the fastest way to spot a routing problem.
• Renew the static addresses on auto-renew. A monitoring address that expires quietly turns your dashboard green for the wrong reason.

How monitoring lies to you