What Is Availability?

Reviewed by Ionut Caval · Updated June 2026

Availability answers a single question: out of all the time in a period, what fraction was the service actually working? It is usually quoted as a percentage in "nines" (99.9%, 99.99%) and is, for most practical purposes, the same idea as uptime. The terms are often used interchangeably, with one nuance: uptime tends to describe what you observe and measure directly, while availability is the broader reliability concept that the formula below makes precise.

The availability formula

In reliability engineering, availability is derived from two component metrics: how often a system fails and how long it takes to recover each time. The standard formula is:

• Availability = MTBF / (MTBF + MTTR)
• MTBF (mean time between failures) is the average time the system runs normally between one failure and the next. Higher is better.
• MTTR (mean time to recovery) is the average time to restore service once a failure begins. Lower is better.

Worked example: a service that runs 200 hours between failures and takes 1 hour to recover is available 200 / (200 + 1), which is 200 / 201, or about 99.5% of the time. Cut the recovery time in half to 30 minutes and availability climbs to 200 / 200.5, roughly 99.75%. The same headline percentage improves either by failing less often (raising MTBF) or by recovering faster (lowering MTTR), which is why both metrics matter.

What the percentages allow

High availability targets sound nearly perfect but still permit real outage time, and each extra nine cuts the allowance by roughly a factor of ten. 99.9% ("three nines") allows about 43 minutes of downtime per month; 99.95% allows about 22 minutes; and 99.99% ("four nines") allows only about 4.3 minutes per month. Because a single extended outage can exhaust an entire monthly budget at once, recovery speed often matters as much as raw stability. For a full breakdown of what each nine permits, see what 99.9% uptime actually means.

Availability is also the metric most often written into contracts. A Service Level Agreement (SLA) commits a provider to a stated availability target over a defined window and sets out remedies, such as service credits, if it is missed. The figure is only trustworthy when measured the way a visitor actually connects, from outside your own infrastructure, since a server can report itself as healthy while DNS, a CDN, an expired certificate, or a broken checkout makes the site unusable. Continuous external uptime and SLA monitoring records the data needed to track availability over time and prove it against an SLA.

See also: Uptime & SLA monitoring