Global HTTP Check

Check whether a website is up and how fast it responds from probes on six continents. See it the way visitors around the world do, not just from where you are.

A site being "down" is almost never a yes-or-no question. It is up from where you are and down from somewhere else, or up everywhere and slow from one continent, or answering fine but returning an error page to every visitor who asks.

This makes one real HTTPS request from probes on six continents and reports what each one got back: the status code, how long the whole response took, and how long until the first byte arrived. Enter a domain to start.

Reading the status codes

A code in the 200s means the server answered normally. The 300s are redirects, which are usually fine and simply mean the address you entered points somewhere else. Both count as healthy here.

The 400s are the server telling you the request was wrong — a 404 for a missing page, a 403 where something is refusing you. That is worth knowing, but it is not an outage. The 500s are the ones to worry about: the server is reachable and running, and its own code is failing.

Total time and first byte are different problems

First byte is how long the server took to start answering. When that number is high, the delay is on the server: a slow database query, a cold cache, an application doing too much work before it says anything at all.

Total is first byte plus the time to send the rest of the response. When first byte is quick but total is not, the server is thinking fast and the response is simply large, or the network path between the probe and the server is the constraint.

What a regional difference actually means

A site that answers in 90 ms from Frankfurt and 800 ms from São Paulo is not broken. It is a single server in Europe with nothing closer to South America, and every visitor in that region pays the same cost. That is the gap a CDN exists to close.

What should bother you is a site that is meant to be behind a CDN and still shows a wide spread. That usually means DNS is handing the same origin address to every region instead of a nearby edge, and the DNS propagation checker will show you whether that is what is happening.

When nothing answers at all

No response is a weaker signal than it looks. It can mean the site is genuinely down, but it can equally mean the name does not resolve, or that a firewall or bot-protection layer is dropping requests it does not recognise — and automated probes get dropped that way routinely.

If every location reports nothing, check the name resolves first with a DNS lookup. A domain whose records are missing or expired looks exactly like a server that is down, and the fix is somewhere completely different.

This is not the same as monitoring

What you get here is one measurement, at one moment, from sixteen places. That answers "is it broken right now", which is the question you have when something looks wrong.

It cannot tell you whether the site was down at three in the morning, or how often it drops out. That needs something watching continuously, which is what the uptime monitor is for. The two answer different questions and neither substitutes for the other.

Frequently asked questions