Is the internet disrupted, or is one website failing?

A failed page load can originate in a device, local network, provider path, domain lookup or application. Record the affected service, time and exact error first. A name-resolution error, connection timeout and server error describe different observations and should lead to different next checks.

Compare independent destinations, including a domestic service, an international service and the affected site. A single failing destination gives you a narrower starting point. Several simultaneous failures suggest a shared dependency, but do not yet identify it. Preserve these observations instead of replacing them with a general statement that the internet is down.

Covered reports what its measurement points observe. A website can respond from one point while another cannot reach it. That difference helps establish the affected scope. A server connected through an operator still does not represent every subscriber or city served by that operator.

Consider a hypothetical case: two measurement points receive errors while several others receive valid responses at the same time. The supported conclusion is that access failed from those two points. A claim that the website is unavailable everywhere, or that a whole country is disconnected, requires much broader evidence.

Current status, incident history and success rate over a period answer different questions. A service may have recovered while its history still shows earlier failures. Missing samples do not establish that the unobserved period was healthy. Compare the same time window and check when the latest measurement arrived.

A successful ping does not prove that sign-in or payment works. A homepage check does not necessarily cover an API. Use the service page to review the tested function, affected points and incident times, then compare with a test from your own connection when assessing your experience.