Every host advertises an uptime number, and they mostly cluster around 99.9%. It sounds close enough to 100% not to matter. The actual math tells a slightly different story.
What 99.9% actually adds up to
99.9% uptime allows for about 8 hours and 45 minutes of downtime over a year. 99.99% —ten times better— allows for about 52 minutes a year. That single extra “9” represents a real difference in infrastructure, not just a marketing rounding error.
Most downtime isn’t dramatic
The image that comes to mind is a big, headline-worthy outage. In practice, most downtime is smaller and more routine: a server restart for maintenance, a resource spike on shared infrastructure, an update that didn’t roll back cleanly. It adds up in minutes here and there rather than one dramatic event.
What to actually check when a host claims a number
- Is there a real SLA, or is the number just marketing copy with nothing backing it up?
- Do you get credited if they miss it, and is that process actually usable?
- Is there a public status history you can check, with real incident explanations?
A number with no way to verify it isn’t worth much more than a guess. We publish our status history in real time for exactly that reason — you shouldn’t have to take an uptime claim on faith.