What the status code means
A code in the 200s means the request succeeded. The 300s are redirects, which are normal but worth understanding -- see below. The 400s mean the request was rejected: 403 is forbidden, 404 is not found, 429 is rate limiting.
The 500s are server errors and are the ones to act on. A 502 or 504 usually means a proxy could not reach the application behind it, which points at the application being down rather than the web server. A 503 often means deliberate maintenance mode or an overloaded service.
Reading the response time
Three requests are made and the spread is reported. The first is often slower than the rest because the TLS handshake and any cold caches are paid once, so compare the minimum against the maximum rather than reading a single figure.
Under 500 ms is good. Between one and three seconds is noticeably sluggish and will affect both conversion and search ranking, since page speed is a ranking signal. Beyond three seconds most visitors abandon the page. Note that this measures server response, not full page load -- images, scripts and fonts add to what a real visitor experiences.
Redirect chains cost time
Every redirect is an extra round trip before the visitor sees anything. One is normal and expected: http to https, or the bare domain to www.
Two or more is worth fixing. The usual cause is chaining rules that each handle one concern -- http to https, then non-www to www, then adding a trailing slash -- when a single rule could go straight to the final URL. On a mobile connection each hop can add hundreds of milliseconds.
What this cannot tell you
A single check from one location at one moment is not uptime monitoring. A site can answer perfectly here and be unreachable from another country, or fail intermittently in a way one request will not catch.
For genuine uptime measurement you need continuous checks from several locations with alerting. This tool answers "is it up right now, from here", which is the right question when you are actively diagnosing something.
Frequently asked questions
Why is the first request slower than the others?
The TLS handshake, DNS resolution and any cold application caches are paid on the first request. Subsequent ones reuse the warmed path, which is why the minimum is the fairer measure of the server's speed.
Does this measure how fast my page loads for visitors?
No. It measures how quickly the server returns the initial response. A visitor also downloads images, CSS, fonts and scripts, which usually dominate the total.
What does a 403 from this tool mean when the site works for me?
Many sites block requests from datacenter addresses or require a browser fingerprint. That protection is working as intended; it is not a fault with your site.
How many redirects are acceptable?
One. Consolidate anything beyond that into a single rule that sends visitors straight to the final URL.