Website Uptime Monitoring: The Complete Guide
Every minute your website is down costs you visitors, sales, and trust. Uptime monitoring is how you find out about those minutes — ideally before your customers do.
What is uptime monitoring?
Uptime monitoring is the practice of automatically checking, on a schedule, whether your website or service is reachable and behaving correctly. A monitoring service sends a request from outside your network every minute or few minutes. If the request fails — or succeeds but returns the wrong thing — you get an alert.
"Uptime" is usually expressed as a percentage over a period. 99.9% uptime sounds great until you do the math: it still allows about 43 minutes of downtime per month. 99.99% allows roughly 4 minutes. Knowing your real number is the first step to improving it.
How a check actually works
A single check is simple: connect, request, evaluate, record.
- Resolve the domain to an IP address.
- Open a connection (and complete a TLS handshake for HTTPS).
- Send the request and wait for a response.
- Decide up or down based on the status code, response time, and content.
- Store the result and, if the state changed, send an alert.
The important nuance is step 4. A server that returns a 500 error is technically "responding" but is not up in any way that matters to a user.
Check types you should know
Different problems need different checks:
- HTTP/HTTPS — the default. Confirms the page loads with a healthy status code (200–399).
- Keyword — fetches the page and confirms a specific word or phrase is present. Catches "the server is up but showing an error page" situations.
- Ping (ICMP) — confirms the host is reachable at the network level. Good for servers, less meaningful for web apps behind a CDN.
- TCP port — confirms a specific port (a database, mail server, game server) accepts connections.
- DNS — confirms your domain resolves to the records you expect.
- SSL certificate — warns you before your certificate expires, which is one of the most common and most avoidable outages.
If you only do one, do an HTTPS check with a keyword. It catches the most failure modes for the least setup.
How often should you check?
More frequent checks mean faster detection but more noise and load. A good default is every 1–5 minutes. To avoid false alarms from a single dropped packet, a good monitor confirms a failure with a second or third check, ideally from a different location, before declaring an outage.
Alerts: the part that actually matters
Monitoring is worthless if the alert doesn't reach you. Set up at least two channels — for example email plus one of SMS, Telegram, or a webhook into Slack. Make sure you'll be notified again if an outage continues, and notified when it recovers.
Getting started
You don't need a complex setup. Pick the most important URL on your site, add an HTTPS check with a keyword, turn on email and one real-time channel, and add an SSL-expiry alert. That covers the majority of real incidents.
[Create a free SiteIsOnline account](/register) and add your first monitor in under a minute — or try the free [website down checker](/tools/website-down-checker) first, no signup required.