A monitoring workstation with several screens beside data center server racks

What Is Uptime? Meaning, Formula and 99.9% Explained

Teckohost earns a commission when you buy through some links on this page, at no extra cost to you. Prices were checked on November 6, 2026.

Uptime is the share of time a website, server or service is up and reachable by visitors. It is shown as a percentage, usually over a month or a year. A host that promises 99.9% uptime allows about 43 minutes and 48 seconds of downtime in an average month.

Below is how to read that number, calculate it and check it yourself.

What is uptime? The share of time your website is reachable

Website uptime measures how long your site answers visitors when they ask for a page. The opposite is downtime, the time your site is offline, returns errors or does not load at all.

Think of it as a ratio. If your site was reachable for 29 days and 23 hours of a 30-day month, your uptime was about 99.86%. Everything else was an outage of some kind.

The word also has a second meaning. On a computer or phone, uptime is how long the device has run since its last restart. This page covers the first meaning: the availability of a website or web server, the one that matters when you choose a hosting provider.

Uptime sits close to what web hosting is in the first place. Your host keeps a web server running around the clock so that anyone can reach your files. Uptime is the score for how well it does that job.

The uptime formula is simple: uptime divided by total time

To calculate uptime, divide the minutes your site was available by the total minutes in the measurement period, then multiply by 100.

For a 30-day month there are 43,200 minutes. If your site was down for 90 minutes in total, the math is (43,200 minus 90) divided by 43,200, times 100. That gives 99.79% uptime.

Two choices change the answer, so write them down before you compare numbers:

  • The measurement period. A host may report uptime measured monthly, while a monitoring tool may report it over a rolling 90 days or a year.
  • The measurement method. One service pings the server. Another loads a full page and checks that the response is correct. The second method catches more problems.

The nines tell you how much downtime is allowed

People describe uptime by counting the nines. Three nines means 99.9%. Four nines means 99.99%. Each extra nine cuts the allowed downtime by a factor of ten, and each one costs the provider more to deliver.

Uptime percentage Common name Downtime per month Downtime per year
99% Two nines About 7 hours 18 minutes About 3 days 15 hours
99.9% Three nines About 43 minutes 48 seconds About 8 hours 46 minutes
99.95% Three and a half nines About 21 minutes 54 seconds About 4 hours 23 minutes
99.99% Four nines About 4 minutes 23 seconds About 52 minutes 34 seconds
99.999% Five nines About 26 seconds About 5 minutes 15 seconds

These figures use an average month of 30.44 days and a year of 365.25 days. They are plain math, not a measurement of any host.

Is 99.9% uptime good? For most small sites, yes

A 99.9% uptime guarantee is the industry standard for shared hosting. For a blog, a portfolio or a local business site, 43 minutes of downtime in a month rarely costs much. Most visitors never notice.

The picture changes for an online store. When checkout is down, you lose sales and you can end up with failed transactions that confuse customers. For an e-commerce site that earns money every hour, four nines is a fair target.

Five nines is a goal for banks, payment networks and emergency systems. It is not something you buy on a standard hosting plan, and a plan that claims it deserves a close read of the terms.

Is 100% uptime possible? Not in practice. Hardware fails, networks drop, software needs updates. A promise of 100% usually comes with fine print that excludes the very events that cause outages.

An uptime guarantee is a promise with conditions

Most hosting providers publish an uptime guarantee, often written as an uptime SLA. SLA stands for service level agreement, the document that sets out what the provider promises and what happens if it misses.

Uptime guarantees are not the same as actual uptime. A guarantee is a target the host commits to. Actual uptime is what really happened, and only your own monitoring or an independent service can tell you that.

Here is what the hosts we link to say about themselves. These are their own claims, checked on November 6, 2026, not our measurements:

  • Bluehost lists a 99.99% uptime SLA on its current plans.
  • SiteGround says its network is backed by a 99.9% uptime guarantee, with automatic failover and proactive monitoring.

We have not run our own uptime tests yet, so we do not rank hosts by uptime. See our Bluehost review and SiteGround review for the plan details we can verify.

Read the fine print before you trust a guarantee

Three details decide whether a guarantee means anything to you.

Planned maintenance often does not count

Most SLAs exclude scheduled maintenance and planned maintenance windows. If your host takes servers offline at 2 a.m. for an upgrade and announces it in advance, that time usually does not count against the guarantee.

The remedy is usually a credit, not cash

If a host misses its uptime SLA, the usual remedy is a service credit, such as a few days of free hosting. It rarely covers the sales you lost. You often have to file a claim within a set number of days and show your evidence.

Partial outages can slip through

A site can be up but broken. Slow database queries, a checkout page that fails or a server that answers with errors can be partial outages. Whether they count depends on how the host defines downtime.

When you compare plans, find the SLA page on the host’s site and read the definition of downtime, the exclusions and the claim process. Our guide on how to choose a web host lists the other checks worth making.

What causes downtime on a website

Downtime comes from a short list of causes. Some sit with your hosting provider and some sit with you.

  • Hardware and data center problems. Failed disks, power cuts and cooling faults at the data center.
  • Network infrastructure. A failed router, a cut cable or a routing error between the host and your visitors.
  • Traffic spikes. A burst of visitors can overwhelm a small plan, especially during peak hours.
  • Software faults. A broken plugin, a bad update or a problem in the operating system or web server.
  • Attacks. Denial-of-service floods and malware can take a site offline. See what website security involves.
  • Expired items. A lapsed domain or an expired SSL certificate can make a healthy site unreachable. Read what an SSL certificate is to avoid the second one.
  • Routine maintenance. Updates and migrations, planned or not.

Your host controls the first three. You control the rest, so keep software updated and set renewal reminders. Our page on why you need reliable hosting explains how the pieces connect.

Server uptime, service uptime and website availability are not the same

Server uptime tells you the machine is running. Service uptime tells you a specific service, such as the database or email, is working. Website availability tells you that users can reach a page and get a correct answer. A web host can report strong server uptime while your site still fails, so check website availability too.

Network conditions between your visitors and the data center also matter. A user in one region may see an outage that users elsewhere do not. This is one reason external checks from several places give a fairer picture than a single report.

Maximum uptime is the goal, but total downtime is what your customers feel. Even a few minutes at the wrong moment, such as during business hours or a sale, can cost more than an hour at 3 a.m.

Track downtime incidents with dates and times, and compare uptime metrics month to month. Many providers will show these on request. Most websites never need more than three nines, but most websites do need to know when they fall short.

Uptime and response time are different problems

A site that takes 20 seconds to load is technically up, but visitors may leave before it appears. Response time, the time the server takes to start answering, is a separate measure of service reliability.

Watch both. Uptime tells you whether visitors can reach you. Response time tells you whether they will wait. A good monitoring tool records each of them, and a good host works on both.

How to monitor uptime yourself

You do not have to take a web host’s word for it. Website monitoring and external services check your site from outside your host’s network, so they are independent of the host’s own reports. Effective monitoring uses external checks, alerts and a record of hosting uptime over time.

  1. Pick an external monitor. Many uptime monitoring services offer free tiers that check a site every few minutes. Compare the check interval and the alert options before you choose.
  2. Check the right page. Monitor your home page and one important page, such as checkout or a contact form. A home page can load while the rest of the site is broken.
  3. Use multiple locations. Checks from more than one region separate a local network problem from a real outage.
  4. Turn on alerts. Email or text alerts let you act within minutes instead of hearing about it from a customer.
  5. Read the uptime reports. Review the monthly report and compare the actual uptime with your host’s guarantee. Keep the records if you ever need to file a claim.
  6. Check the status pages. Hosts publish status pages with incident notes. These help you tell a host-wide outage from a problem on your own site.

If your own number and your host’s number disagree, ask which measurement method each one uses. A check every five minutes can miss a short outage. A host may also exclude events your monitor counts as downtime.

How to pick a host with dependable uptime

Use these checks in order:

  1. Find the written uptime SLA and read what counts as downtime.
  2. Look for a status page and a history of past incidents.
  3. Check how the host handles failover, backups and monitoring.
  4. Match the plan to your traffic, so a spike does not push you offline.
  5. Set up your own external monitoring from the first day.

Cloud and managed plans often spread your site across more servers, which can lower the risk of a single failure. Read what cloud hosting is and shared hosting vs VPS to see what each tier offers. Our best web hosting page collects the providers we cover.

Uptime questions people ask

What does 99.9% uptime mean?

It means the service is available 99.9% of the time. Over an average month, that allows about 43 minutes and 48 seconds of downtime. Over a year, it allows about 8 hours and 46 minutes.

What is a good uptime for a business website?

For most small business sites, 99.9% is a reasonable minimum. Stores that earn money every hour should aim higher, around 99.99%, and keep their own monitoring running.

Does planned maintenance count as downtime?

Often not. Many SLAs exclude scheduled maintenance. Check the exclusions in your host’s agreement, because the same outage can count under one provider’s terms and not another’s.

How do I check the uptime of a website?

Use an external uptime monitoring service. It requests your page on a schedule, records every failure and sends alerts. Many offer a free plan that is enough for one site.

Can I get a refund when a host misses its uptime guarantee?

Usually you get a service credit, not a refund. Terms differ by host, so read the claim process. Money-back guarantees are separate and apply to the first days after you buy.

Share your love
Tumisang Bogwasi
Tumisang Bogwasi