Uptime monitoring is the practice of automatically checking, from outside your own infrastructure, whether your website or service is reachable and responding correctly — and alerting you immediately when it isn’t. It’s the smoke detector of running anything on the internet: cheap, simple, and the difference between a 4-minute incident and a 4-hour one.
How uptime monitoring works
A monitoring service runs checks against your site on a schedule from servers in multiple regions:
- Request. Every 30 seconds to 5 minutes, a checker sends a request — usually HTTP(S) — to your URL.
- Evaluation. It verifies the response: status code (200 vs 500), response time, and optionally page content or certificate validity.
- Confirmation. On failure, good monitors re-test from a second location to rule out local network noise and avoid false alarms.
- Alert. A confirmed failure triggers notifications — email, Slack, SMS, phone — typically within a minute.
- Recovery + reporting. The monitor logs the outage duration, notifies you on recovery, and rolls everything into an uptime percentage over time.
That percentage is the metric everyone quotes. The gap between the “nines” is bigger than it looks:
| Uptime | Downtime allowed per month | Per year |
|---|---|---|
| 99% | ~7.3 hours | ~3.7 days |
| 99.9% | ~43 minutes | ~8.8 hours |
| 99.99% | ~4.4 minutes | ~53 minutes |
| 99.999% | ~26 seconds | ~5.3 minutes |
(For how uptime relates to SLAs, SLOs and reliability targets, see uptime vs availability vs reliability.)
The main types of checks
- HTTP/HTTPS checks — the workhorse. Request a URL, expect a healthy status code. Catches most outages.
- Keyword/content checks — verify the page actually contains expected text. Catches the nastiest failure mode: a page that returns 200 but renders an error or a blank screen.
- SSL certificate checks — alert before certificates expire. An expired cert is self-inflicted downtime with a countdown you could have watched.
- Ping (ICMP) and port checks — is the host alive, is port 443/25/5432 accepting connections? Useful for servers, mail and databases.
- API checks — call an endpoint, validate status and response body. Essential when your product is an API.
- Heartbeat (cron) checks — reversed logic: your scheduled job pings the monitor; if the ping doesn’t arrive, you’re alerted. The only way to notice a silently dead backup job.
- Transaction/browser checks — scripted multi-step journeys (login, checkout) in a real browser. This is the border where uptime monitoring becomes synthetic monitoring.
What uptime monitoring is not
- Not server monitoring. External checks can’t see CPU, memory or disk — that’s inside-out server monitoring, the natural companion.
- Not real user monitoring. RUM measures actual visitors’ experience; uptime checks are simulated visits at fixed intervals.
- Not a status page — though most tools bundle one, and you should use it: telling users about an incident before they discover it themselves is half of incident handling done well.
Why it matters (even for small sites)
Downtime is invisible from the inside. Your site can be down for hours while everything looks fine from your own network — DNS issues, regional outages and expired certificates are all failures your own browser cache and location can hide. External monitoring is the only honest witness. It also builds the historical record — uptime percentages, response-time trends, outage logs — you’ll want when choosing hosting, debugging a flaky provider, or backing up an SLA claim.
How to get started (5 minutes, $0)
- Pick a tool with a free tier — hosted (Better Stack, UptimeRobot) or self-hosted (Uptime Kuma). Our free uptime monitoring comparison covers the honest limits of each.
- Add an HTTPS check for your homepage and your most important endpoint.
- Add an SSL-expiry check and a keyword check on a page that exercises your database.
- Route alerts somewhere you’ll actually see them (Slack/phone, not a dead inbox).
- When downtime starts costing money, upgrade to 30–60-second checks and phone alerts — the full walkthrough is in how to monitor website uptime, and the paid landscape is in our best uptime monitoring tools guide.