Web Hosting Uptime: What It Means and How to Check Yours
A 99.9% uptime guarantee sounds like your site is always on. Do the maths and it permits just under nine hours of downtime every year. That figure is on the host's terms, measured the way they choose, logged when they decide something counts as 'down'. Whether any of that downtime lands at 2am or right in the middle of your busiest trading hour is not covered. Here is what those numbers actually mean, and how to find out what your host is really delivering.
On this page
What 99.9% Uptime Actually Adds Up To
The maths is simple once you run it. 99.9% uptime across a year leaves 0.1% unaccounted for. That is 8 hours and 45 minutes. Step up to 99.99% and you are still looking at around 52 minutes a year where your site can legitimately go dark under the terms of the SLA.
Neither figure is dishonest. The problem is that most people read the percentage and stop there. Fifty-two minutes spread across a year sounds trivial. One unannounced outage on a Friday afternoon during a product launch is a different matter entirely, and the SLA treats both the same way.
Why the Percentage Is Only Half the Story
Timing is everything. A 15-minute outage at 2am barely registers. The same outage at noon on a Monday, when a Googlebot crawl is midway through your site, or when someone has just clicked a paid ad, costs you something real.
Most hosting SLAs say nothing about when downtime is scheduled or when it tends to cluster. Maintenance windows, server migrations and resource spikes follow the host’s schedule, not yours. Check your own traffic data in Google Analytics or Search Console. If your peak hours are 10am to 2pm, that is the window you need reliable uptime in, not just a good annual average.
How Hosts Measure Uptime (And What They Leave Out)
Host-reported uptime almost always measures one thing, whether the server responds to a ping. If the server sends back any response, it counts as up. That is not the same as your site working properly.
A page can respond to a ping and still be returning a database timeout to every real visitor. Slow TTFB, PHP errors, a broken connection to your database, a failed CDN handoff , none of these necessarily trigger a host’s own monitoring as a downtime event. Partial failures like these are invisible to most SLA calculations but very visible to your users and to search engine crawlers.
This is why host-reported uptime figures almost always look better than independent monitoring shows. The host is measuring their infrastructure. You need to measure what visitors actually see.
How to Check Your Host’s Uptime Yourself
Two tools worth knowing are UptimeRobot and Better Uptime. Both offer free tiers that handle basic HTTP monitoring. Set them up to check your homepage URL, not just your domain, because that tests the full stack rather than just a DNS response.
Check every five minutes rather than every minute. One-minute polling on a free plan can generate noise from brief network blips that are not genuine outages. Five-minute intervals give you a clean picture without false positives. Run it for four to six weeks before drawing conclusions, then compare the logged availability percentage against what your host is claiming. Any consistent gap between the two is worth questioning.
For a deeper check on how your hosting affects real page load performance, the fixes that move Core Web Vitals scores cover TTFB and server response time in detail.
What Downtime Does to Your SEO
Google does not penalise a site for one isolated outage. Recurring short outages are a different matter. If Googlebot visits your site during a crawl window and repeatedly hits a 503 or a timeout, it scales back how often it crawls. That wastes crawl budget and delays indexing for new or updated pages.
The signal is quiet, which is what makes it easy to miss. You will not see a ranking drop you can point to directly. What you will see is pages taking longer than expected to appear in search results, or re-crawl rates dropping off in Search Console. Reliability is a background quality signal. It does not move rankings on its own, but consistent unreliability slowly erodes the conditions that good rankings depend on.
When It’s Time to Move Hosts
Persistent outages during peak hours, a consistent gap between your host’s claimed uptime and what independent monitoring shows, or repeated partial failures that never appear in any SLA report are all legitimate reasons to consider moving.
The honest caveat is this, a poorly planned migration can cause more disruption than the original problem. DNS propagation, database transfer errors and misconfigured redirects can take a site offline for longer than any individual hosting outage. If you do decide to move, do it methodically. Test the new environment before switching DNS, understand what actually matters in a hosting setup before committing, and always keep a full backup you can restore from if something goes wrong during the transfer.
A better host will not fix SEO problems on its own. But a reliable one stops your hosting from quietly working against everything else you are doing right.