Availability
Availability describes whether a website, server, database, or other system is reachable and working right now. It is a closely related but distinct concept from uptime: availability is a point-in-time state ("is it up at this moment?"), while uptime is the historical percentage of time a system has been available over a defined period, such as a month or a year.
How availability is measured
Because every internet-connected system depends on complex infrastructure — servers, networks, DNS, upstream providers — nothing is available 100% of the time in practice. Availability is measured by repeatedly checking a system from the outside at a fixed interval and recording whether each check succeeded or failed. A single check is a snapshot; a series of checks over time is what produces a meaningful uptime percentage, and checks confirmed from multiple independent locations avoid mistaking one network's local problem for a real outage.
What affects availability
Availability problems come from many directions: server or hardware failure, an overloaded application unable to keep up with traffic, a software bug or bad deployment, a network or routing failure between the checking location and the server, DNS misconfiguration, or a certificate expiring and breaking HTTPS connections. No single cause dominates — which is exactly why monitoring needs to check multiple layers (network reachability, the application itself, DNS, certificates) rather than assuming one green check covers everything.
What good availability looks like
Availability targets are usually expressed the same way uptime is: 99.9% ("three nines") corresponds to about 8 hours 45 minutes of unavailability per year, and 99.99% ("four nines") to about 52 minutes 34 seconds. See the full percentage-to-downtime breakdown on the uptime page. Even at 99%, a system is still unavailable for the equivalent of several days a year — usually enough to be noticed and to cost real business if it happens in one stretch.
How HostTracker tracks availability
HostTracker checks availability from 300+ monitoring locations worldwide, using any combination of 14 check types — HTTP/S, ping, TCP port, transaction, database, and more — so unavailability is caught and confirmed from outside your own infrastructure, independent of what a hosting provider's own status page reports. Run a free check with the website down checker, or set up continuous distributed availability monitoring so you're alerted the moment something changes. See why websites need monitoring for the fuller case.