When deciding the level of availability we want for our services, the target that we want to achieve is often described as a percentage of time the service is available.
It's worth noting that 100% availability, while it sounds like a nice thing to have, is immensely expensive to achieve, and also pointless to a big extent - what's the point of having 100% available service which is used on a below-100% available device, through a below-100% available network?
This number is famous as a "number of nines". A system that's 99% available is said to have "two nines" service availability target, while 99.999% available system has "five nines".
The table below represents availability targets, together with how much downtime they allow per year/quarter/month/week/day:
Availability | Year | Quarter | Month | Week | Day |
---|---|---|---|---|---|
90% | 36.5d | 9d | 3d | 16.8h | 2.4h |
95% | 18.25d | 4.5d | 1.5d | 8.4h | 1.2h |
99% | 3.65d | 21.6h | 7.2h | 1.68h | 14.4m |
99.5% | 1.83d | 10.8h | 3.6h | 50.4m | 7.20m |
99.9% | 8.76h | 2.16h | 43.2m | 10.1m | 1.44m |
9.95% | 4.38h | 1.08h | 21.6m | 5.04m | 43.2s |
99,99% | 52.6m | 12.96m | 4.32m | 60.5s | 8.64s |
99.999% | 5.26m | 1.3m | 25.9s | 6.05s | 0.87s |
Status: #🌲 References: