# High Availability (HA)

> A service built to stay up nearly all the time. Each part has a spare copy.

**Field:** Infrastructure  
**Category:** Reliability  
**Also known as:** HA, Highly Available, Availability, Nines of Availability

## Think of it like

It is like a plane with two engines. If one stops, the other gets you home.

## Definition

High availability means a service stays up almost all the time. It is built by running spare copies of each part, so no one failure stops it. When one copy fails, traffic moves to another, often in seconds. Availability is counted in nines. 99.9% is three nines, and 99.99% is four.

## Example

Three nines, 99.9%, allows about 8.8 hours down a year. Four nines, 99.99%, allows under an hour, and costs far more to reach.

## Why it matters

For a shop or a bank, each minute down is lost money and lost trust.

## Common mistake

Running two copies on one machine or in one data center. One power cut then takes both down.

## Three principles

- **No single point.** Every part has a spare.
- **Fast switch.** Traffic moves to a healthy copy.
- **Counted in nines.** Each extra nine costs much more.

## Related terms

- [Fault Tolerance](https://learn.benibix.com/term/infrastructure/fault-tolerance)
- [Health Check](https://learn.benibix.com/term/infrastructure/health-check)
- [Load Balancer](https://learn.benibix.com/term/infrastructure/load-balancer)
- [Service Level Agreement (SLA)](https://learn.benibix.com/term/infrastructure/service-level-agreement)
- [Single Point of Failure](https://learn.benibix.com/term/development/single-point-of-failure)
- [Availability Zone](https://learn.benibix.com/term/infrastructure/availability-zone)
- [Disaster Recovery](https://learn.benibix.com/term/infrastructure/disaster-recovery)

## Told apart from

- [High Availability vs Fault Tolerance](https://learn.benibix.com/compare/high-availability-vs-fault-tolerance): People use the two names as one idea, and they are not. With high availability, a service may blip for a few seconds while traffic moves to a spare. With fault tolerance, spares run at the same time, so there is no blip at all. The second costs far more, so most sites only need the first.

![A load balancer at the top sends traffic down to three servers. The middle server is drawn dashed, and its health check says fail. The load balancer has stopped sending it traffic, and the two other servers carry on serving users.](https://learn.benibix.com/diagrams/inf-high-availability.svg)

---

Source: https://learn.benibix.com/term/infrastructure/high-availability
Author: Behnia Zarei (a developer in Vienna, Austria, who builds and teaches data, AI and business intelligence at BENIBIX)
Publisher: BENIBIX Learn (https://learn.benibix.com/)
Last updated: 2026-11-19
Licence: CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). Quote freely with attribution.
