Availability
Availability is the share of time, or of requests, for which a service can be used when someone needs it.
Uptime and availability are often used as the same thing. When they are told apart, uptime is whether the system is running, and availability is whether users can actually use it: a server can be up while a broken load balancer makes the service unavailable.
Availability can be measured by time (minutes the service worked) or by requests (the share of requests that succeeded). Request-based availability weights busy hours more, which is closer to how many users were affected.
In reliability engineering it is also estimated from how often a system fails and how long repairs take, using MTBF and MTTR.
At a glance
availability = MTBF ÷ (MTBF + MTTR) example: 720 h ÷ (720 h + 0.5 h) ≈ 99.93%
Calculate availability from MTBF and MTTR →
Questions
Is availability the same as uptime?
Often they are used as synonyms. Strictly, uptime says the system was running and availability says users could use it, which can differ when a dependency in front of the system fails.
How does redundancy change availability?
When any one of several copies is enough, their failure rates multiply: two independent copies at 99% give 99.99%. When all parts are needed, their availabilities multiply instead and the total goes down.
Related terms
Monitor it with OpenUptime
Checks from one or several regions, alerts to the tools you use, and a status page for your customers. Free to start, open source under the MIT license.
可用性（Availability）
可用性是指在用户需要时，服务能够正常使用的时间（或请求）所占的比例。
Uptime 和可用性常被当作同一个意思。区分时，uptime 指系统是否在运行，可用性指用户能否真正使用它：服务器可能在运行，但负载均衡器出了问题，服务对用户来说仍然不可用。
可用性可以按时间测量（服务正常的分钟数），也可以按请求测量（成功请求的比例）。按请求计算会给繁忙时段更大的权重，更接近实际受影响的用户数。
在可靠性工程中，它也可以根据系统多久出一次故障、修复要多久来估算，也就是用 MTBF 和 MTTR。
速览
可用性 = MTBF ÷（MTBF + MTTR） 例如：720 小时 ÷（720 小时 + 0.5 小时）≈ 99.93%
常见问题
可用性和 uptime 是一回事吗？
常被当作同义词。严格来说，uptime 表示系统在运行，可用性表示用户能用上，当系统前面的某个依赖出故障时，两者就会不同。
冗余如何影响可用性？
当多个副本中只要一个正常就行时，故障率相乘：两个独立的 99% 副本合起来是 99.99%。当每个部分都必须正常时，则是可用性相乘，整体会下降。