Uptime is the proportion of time a website or server is available and working, usually given as a percentage over a month or a year. Its opposite, downtime, is any period when visitors cannot load the site or get an error instead.
How uptime works
Hosts promise uptime in a service level agreement, often 99.9 percent. The figures sound alike, but the difference adds up. Over a 30-day month, 99.9 percent still allows about 43 minutes of downtime, 99.5 percent allows about three and a half hours, and 99 percent allows over seven hours. The guarantee usually means a partial refund of hosting fees if it is missed, not compensation for lost business, and planned maintenance is often excluded.
Uptime is measured by monitoring services that request your site from several locations every minute or few minutes. If a check fails, perhaps a timeout or a 5xx server error, the service tries again from elsewhere to rule out a local glitch, then alerts you by email, text or app notification. Better monitors also check for a word that should appear on the page, because a server can answer with a “200 OK” while showing a blank screen or a database error.
Downtime has many causes beyond the host: an expired domain or SSL certificate, a DNS change gone wrong, a plugin update that crashes the site, a traffic spike on shared hosting, or an attack.
Why it matters
The direct cost is lost enquiries and sales. A wedding venue in the Cotswolds that goes down on a Sunday evening, when couples are browsing together, loses exactly the visits it spent money attracting. Paid advertising keeps sending clicks to the broken page unless someone notices and pauses the campaigns.
The search cost depends on how long the problem lasts. Google copes with brief outages: if Googlebot meets server errors, it slows its crawling and tries again later. Errors that persist for more than a day or two are different. Google starts to drop affected URLs from its index, and it can take time to crawl them back after the site recovers. Repeated instability also reduces how much Google is willing to crawl, which matters for larger sites managing their crawl budget.
Common mistakes
- Finding out from a customer. Without monitoring, downtime overnight or at weekends can last hours before anyone knows.
- Trusting the host’s status page alone. It shows the host’s network, not whether your particular site is loading.
- Checking only the homepage. The homepage may be cached and fine while the booking page or checkout returns errors.
- Using the wrong status code for maintenance. Planned work should return a 503 with a Retry-After header, which tells Google to come back later. A maintenance page served as a normal 200 risks being indexed in place of your content.
- Letting the domain or certificate lapse. These are among the most avoidable causes of downtime.
How to act on it
Set up an external uptime monitor; several reputable services have free tiers that check every few minutes. Monitor the homepage and your most valuable page, such as contact or checkout, and send alerts to at least two people. Add expiry reminders for your domain and certificate, and keep a recent backup so a broken update can be rolled back quickly.
Review the monitor’s monthly report. If downtime keeps recurring, ask your host for the cause, and consider managed hosting if the answer is resource limits. Checking server errors and crawl stats in Search Console is part of any technical SEO review I run, because a site that is often unavailable undermines every other improvement.