Most businesses find out their website is down when a customer mentions it, which means it has usually been down for hours. Monitoring costs almost nothing and changes that entirely.
Uptime monitoring checks your site at regular intervals and alerts you when it stops responding. Monitor more than the homepage: check a page that hits the database, your contact form, and your checkout if you have one. Expect 99.9 percent uptime from decent hosting, which still allows around 43 minutes of downtime a month.
What to monitor
- Homepage. The baseline check.
- A page that queries the database. A static homepage can serve from cache while the database is down.
- Your contact form endpoint. Forms fail independently of the site, and silently.
- Checkout, if you sell.
- SSL certificate expiry. An expired certificate blocks the site entirely and is entirely preventable.
- Domain expiry. Rarer and more catastrophic.
- Response time, so you see degradation before an outage.
Monitoring only the homepage is the common mistake. A cached homepage responding fine while the database is unreachable means every form on the site is failing and the monitor is green.
Realistic uptime expectations
| Uptime | Downtime per month | Typical of |
|---|---|---|
| 99 percent | About 7 hours | Poor hosting |
| 99.9 percent | About 43 minutes | Decent shared hosting |
| 99.95 percent | About 22 minutes | Good managed hosting |
| 99.99 percent | About 4 minutes | Enterprise infrastructure |
Nobody achieves 100 percent. What matters more than the headline figure is how quickly problems are detected and how quickly someone responds.
Setting it up sensibly
- Check every one to five minutes. Hourly checks mean an outage can run most of an hour unnoticed.
- Require two consecutive failures before alerting, to avoid noise from transient blips.
- Check from more than one location, so a regional network issue is not read as an outage.
- Alert through more than one channel. Email alone fails when the problem is email.
- Decide who responds, and what they do. An alert nobody owns is a notification, not monitoring.
Beyond uptime
Monitoring that a page returns a 200 status does not mean it is working. A page can respond successfully while displaying an error, or while the form on it silently fails.
Content checks solve this: assert that a specific phrase appears on the page. If your homepage returns 200 but no longer contains your company name, something is badly wrong and a simple status check would miss it.
Frequently asked questions
What is uptime monitoring?
A service that checks your website at regular intervals and alerts you when it stops responding. It exists so you find out about outages from a system rather than from a customer.
How much does uptime monitoring cost?
Free tiers cover basic checks on a small number of URLs and are enough for many small businesses. Paid plans typically run from a few dollars a month and add more frequent checks, more locations and better alerting.
What is a good uptime percentage?
99.9 percent is a reasonable expectation from decent hosting, which still permits around 43 minutes of downtime a month. Anything below 99 percent suggests a hosting problem worth solving.
Should I monitor more than my homepage?
Yes. A cached homepage can respond normally while the database is down, meaning every form is failing while your monitor shows green. Check a database-driven page, your form endpoint and your checkout.
What should I do when I get an alert?
Confirm it is real by loading the site yourself, check whether your host has reported an incident, and check whether it is a certificate or DNS problem rather than the server. Agree in advance who does this, since an alert nobody owns achieves nothing.