What is Uptime?
Uptime is the percentage of time that a server or website is up and reachable over a measurement period, usually a calendar month or year. The remaining time is downtime. Web hosts commonly advertise a 99.9% uptime guarantee, which still allows roughly 43 minutes of downtime in a 30-day month.
More About Uptime
A high uptime percentage means a server was down for only a small slice of the measurement period, not that it never goes down. Even 99.9% uptime allows roughly 8 hours and 46 minutes of downtime per year. Those interruptions have a cost: while your site is down, visitors can’t reach it, sales and signups stop, and repeated outages erode trust. That’s why uptime is one of the first specs to check when choosing a host.
Some web hosts keep backup servers ready to take over when a server fails. Server clustering is another strategy: a cluster is a set of interconnected servers that share the same workload, so when one fails, the others absorb its share. A load balancer ties the cluster together, routing visitors away from a failed server automatically.
How uptime is measured
Uptime is calculated as a percentage of a measurement window, usually a calendar month: the time the service was up, divided by the total time in the period. In practice, an uptime monitor requests your site at a set interval and logs every failed check as downtime, so the measured figure depends on how often the monitor checks and on what counts as down. UptimeRobot’s uptime guide (updated February 2, 2026) notes that planned maintenance, partial outages, and slow responses are often excluded or disputed after the fact. Don’t rely on your host’s own reporting alone: run a free third-party uptime monitor so you have independent numbers when it matters.
What 99%, 99.9%, and 99.99% actually allow
Small-looking decimals hide big differences, and each extra nine cuts the allowed downtime by a factor of 10. Assuming round-the-clock operation, the SLA calculator uptime.is puts the allowances at:
- 99% uptime: about 7 hours 18 minutes of downtime per month, or 3 days 16 hours per year
- 99.9% uptime: about 43 minutes 50 seconds per month, or 8 hours 46 minutes per year
- 99.99% uptime: about 4 minutes 23 seconds per month, or 52 minutes 36 seconds per year
Treat 99.9% as the practical floor for a production website. If downtime costs you money directly, look for 99.99% or a credit-backed 100% guarantee.
Uptime vs. availability vs. reliability
The three terms travel together, but they measure different things:
- Uptime: the percentage of a measurement period that a server was up and reachable. It caps total downtime but says nothing about how that downtime was spread out.
- Availability: often used interchangeably with uptime, but strictly it asks whether the service works when users try it. A server can be up while the application on it returns errors.
- Reliability: how often failures occur and how long each one lasts. A single 8-hour outage and 90 outages of 5 minutes each both fit inside a year of 99.9% uptime, but they are very different experiences for visitors.
Two hosts can report identical uptime with very different outage patterns. So don’t stop at the headline percentage: check how often outages happen and how long they last. Many hosts publish that history on a public status page.
Uptime guarantees and SLAs
Many web hosts advertise a 99.9% uptime guarantee with their hosting plans. A guarantee is a commitment, not a probability: if availability falls below the promised figure, the host owes you compensation, usually service credits. The details live in the service level agreement (SLA), which defines how uptime is measured, what downtime is excluded, and what credit you get when the target is missed.
A 99% guarantee still allows about 3 days and 16 hours of downtime per year, so check the number and the remedy behind the word “guarantee”, not just the word itself. DreamHost’s Terms of Service (updated June 22, 2026) go further and guarantee 100% uptime, crediting your account a day of hosting for each hour of interruption. Whoever you host with, read the SLA before you buy, not after an outage.
Frequently Asked Questions
Is 100% uptime actually possible?
No host avoids downtime entirely; hardware fails and networks hiccup. A 100% uptime guarantee is a credit-backed promise, not a physics claim: the host compensates you when downtime happens rather than promising it never will.
Does scheduled maintenance count against uptime?
Usually not. SLAs typically exclude scheduled maintenance from the uptime calculation, so a month that included a planned maintenance window can still report as 100% uptime. That's also why your own monitor may show lower uptime than your host reports: it counts every minute your site was unreachable.
Do slow responses count as downtime?
Usually not. A standard uptime check tests whether your site responds, not how fast, so a page that loads painfully slowly still counts as up. That makes slow responses a common dispute in SLA claims. If speed matters, monitor response time as a separate metric.
What causes website downtime?
Common causes include hardware failure, network outages, traffic spikes beyond your plan's capacity, DDoS attacks, and botched software updates. Redundancy across independent hardware and network paths can reduce the first two risks; capacity planning, DDoS protection, and testing updates on a staging site can reduce the others.
