Availability is the measure of a system's ability to be operational and accessible when users need it, expressed as a percentage of total time. It is closely related to uptime but can be defined more precisely — for example, a service might be "up" but returning errors, which would count as unavailable even though the server is running.
Availability is calculated as: Availability = (Total Time - Downtime) / Total Time × 100%. It can also be expressed using MTBF and MTTR: Availability = MTBF / (MTBF + MTTR). This second formula highlights that availability depends on both how often failures happen and how quickly they are resolved.
Different levels of availability map to different amounts of allowable downtime: 99% allows 3.65 days/year, 99.9% allows 8.77 hours/year, 99.99% allows 52.6 minutes/year, and 99.999% allows only 5.26 minutes/year. Teams choose their target based on business requirements, user expectations, and engineering investment capacity.
