Website Uptime Explained: SLAs, Monitoring and What 99.9% Really Means
What uptime percentages mean in minutes of downtime, how hosting SLAs really work, and how to monitor your own site so you know about outages first.
Almost every host advertises an uptime figure such as 99.9%. It sounds close to perfect, but the number only means something when you know how much downtime it allows, how the host measures it and what you get if the promise is broken. This guide translates uptime percentages into real time and shows how to check your site’s availability yourself.
- 99.9% uptime still allows about 44 minutes of downtime in an average month.
- An SLA usually offers service credits, not compensation for lost sales, and often excludes planned maintenance.
- Independent monitoring tells you about outages before your visitors do.
What the percentages mean in practice
An average month has about 730 hours. Here is what common uptime figures allow:
| Uptime | Downtime per month (about) | Downtime per year (about) |
|---|---|---|
| 99% | 7.3 hours | 3.7 days |
| 99.5% | 3.7 hours | 1.8 days |
| 99.9% | 44 minutes | 8.8 hours |
| 99.95% | 22 minutes | 4.4 hours |
| 99.99% | 4.4 minutes | 53 minutes |
Each extra “nine” cuts allowed downtime by a factor of ten, and each one is much harder and more expensive to deliver. Reaching 99.99% generally requires redundant servers, automatic failover and careful operations.
What a service level agreement (SLA) really promises
An SLA is the part of the hosting contract that defines the uptime commitment. When you read it, look for:
- The remedy: usually a credit of a percentage of that month’s hosting fee, which you must request within a set time. It rarely covers lost revenue.
- What counts as downtime: some SLAs count only a complete outage of the server or network, not slow responses or errors in your site’s application.
- Exclusions: scheduled maintenance, attacks, problems caused by your code or plugins, and issues outside the host’s network are commonly excluded.
- How it is measured: by the host’s own monitoring, or by evidence you provide.
A “100% uptime guarantee” is a marketing statement about credits, not a technical promise that nothing will ever fail.
Uptime is not the same as availability for your visitors
Your site can be “up” in the host’s eyes while visitors see problems:
- The server answers, but WordPress returns a database connection error.
- The home page loads from cache, but checkout times out.
- An expired domain or SSL certificate blocks visitors even though the server is fine.
- A DNS problem makes the site unreachable from some networks.
That is why your own monitoring should test what visitors actually experience.
How to monitor your site
- HTTP checks: an external service requests your page every one to five minutes from several locations and alerts you by email, SMS or messaging app when it fails. Many services offer free plans for a few checks.
- Keyword checks: confirm that the page contains a word that should be there, so an error page does not count as “up”.
- Critical paths: monitor the login page, search or checkout, not only the home page.
- Certificate and domain expiry alerts: get warnings weeks in advance.
- A public status page: useful for businesses whose customers need to know whether a problem is on their side.
Keep a simple log of incidents: start time, end time, cause and fix. Over a few months it shows whether your host meets its promises and gives you evidence for SLA claims.
How to reduce downtime
- Choose a host with a clear SLA, a public status page and a history of honest incident reports.
- Keep backups off the server so you can restore elsewhere if the host has a serious failure.
- Use a CDN that can serve cached pages if your server briefly goes down.
- Test updates on a staging site, because many “outages” are caused by a broken plugin or theme update.
- For critical sites, consider hosting with redundancy across servers or data centres.
Frequently asked questions
Is 99.9% uptime good enough?
For most small and medium sites, yes. Stores and services that lose significant money per minute of downtime should look for higher commitments and redundancy.
How do I claim an SLA credit?
Usually by opening a support ticket within the period stated in the SLA, with the dates and times of the outage. Your monitoring log is the best evidence.
Why did my monitor report downtime that my host says did not happen?
The problem may have been in your application, DNS or a network route rather than the server. Check the monitor’s error details and locations.
