What Is Uptime and Why Does It Matter for Websites?
Uptime is the percentage of time a website or service is operational and reachable by users. It matters because every minute of downtime can mean lost visitors, failed transactions, and damaged trust. This explainer covers how uptime is calculated, what causes downtime, how monitoring works, and how to track and improve uptime for your own site.
What uptime actually measures
Uptime is usually expressed as a percentage over a given period (often a month or a year). It answers a simple question: for what share of the time was the service available and responding correctly?
A service that is "up" is not just powered on — it is reachable over the network and returning the expected response. If a server is running but the web page times out, most monitoring tools count that as downtime.
How uptime is calculated
The basic formula is:
uptime % = (total time - downtime) / total time × 100
The difference between "three nines" and "four nines" is larger than it looks. Here is the allowed downtime per common uptime target:
| Uptime target | Downtime per day | Downtime per month (30 days) | Downtime per year |
|---|---|---|---|
| 99% | ~14.4 min | ~7.2 hours | ~3.65 days |
| 99.9% | ~1.44 min | ~43.2 min | ~8.76 hours |
| 99.99% | ~8.6 sec | ~4.32 min | ~52.6 min |
| 99.999% | ~0.86 sec | ~26 sec | ~5.26 min |
This is why 99.9% and 99.99% are treated as very different commitments: the gap is roughly a factor of ten in allowed downtime.
Common causes of downtime
Downtime rarely has a single cause. Typical sources include:
- Outages — hardware failure, power loss, or a cloud provider incident.
- Maintenance — planned upgrades, patches, or migrations that take the service offline.
- Network issues — DNS failures, routing problems, or connectivity loss between users and the server.
- Application errors — crashes, bad deployments, or database problems that make the site unreachable or return errors.
- Traffic overload — a spike that exceeds capacity and slows or blocks responses.
Planned maintenance is sometimes excluded from uptime calculations, but only if it is announced and agreed with users. Unplanned downtime is what most uptime metrics focus on.
How uptime monitoring works
An uptime monitor checks your site or service at regular intervals from one or more locations. Each check sends a request and records whether the response arrived and matched expectations.
What monitors can check:
- HTTP/HTTPS availability — whether the page loads and returns a success status.
- Keyword monitoring — whether a specific word or phrase appears in the response, catching cases where the page loads but shows an error message.
- SSL certificate monitoring — whether the certificate is valid and not close to expiry.
- Cron and heartbeat monitoring — whether scheduled jobs run on time.
When a check fails, the monitor can send an alert by email, SMS, Slack, or other channels. Alerts signal that a check failed — they do not by themselves explain why. The value is in fast notification so you can investigate before users report the problem.
UptimeRobot, for example, offers these monitoring types and states you can start monitoring in 30 seconds, with notifications via email, SMS, Slack, and more. Its published description also mentions 50 monitors available for free. Check the current pricing page for the latest plan details, since limits and features can change.
Practical steps to track and improve uptime
- Set a target. Decide what uptime level your users actually need. An internal tool may be fine at 99%, while a payment page may need 99.99%.
- Monitor from outside. Use an external monitor so you detect problems even when your own network is fine.
- Check more than the homepage. Monitor key pages, APIs, and login flows — the parts users depend on.
- Add keyword checks. A page can return HTTP 200 while showing "Service unavailable." Keyword monitoring catches that.
- Watch SSL expiry. An expired certificate makes the site unreachable for many users even if the server is healthy.
- Route alerts to the right place. Send alerts to a channel someone actually watches, and define who responds.
- Review incidents. After downtime, note the cause and whether monitoring caught it early. Adjust checks and thresholds accordingly.
- Reduce single points of failure. Redundancy, backups, and capacity planning lower the chance and length of outages.
Why uptime matters
Uptime is a direct proxy for reliability. For a business site, downtime can mean lost sales and support load. For a service, it can mean broken integrations and churn. Tracking uptime turns reliability from a vague feeling into a number you can set targets against, alert on, and improve over time.