Availability (the Nines)
Availability is the fraction of time a service is up, usually quoted in 'nines': 99.9% (three nines) is about 43 minutes of downtime a month, 99.99% (four nines) about 4 minutes.
also: nines · 99.9% · three nines · four nines · uptime · high availability
Each nine you add cuts the allowed downtime by roughly ten. Two nines (99%) leaves about 7 hours a month. Three nines (99.9%) leaves about 43 minutes. Four nines (99.99%) leaves about 4 minutes, which is less time than many teams take to notice an alert, acknowledge it, and open the right dashboard.
That is why the cost of availability is not linear. The first three nines come from sensible engineering: health checks, retries, a load balancer, automated deploys. The fourth and fifth nines require removing humans from the recovery path, multi-zone or multi-region redundancy, and tested failover, which is a different order of investment.
Quote availability with its window. 99.9% 'over 30 days' and 99.9% 'over a year' permit very different single outages, and a target with no window attached is not really a target.
faq
Questions & answers
- How much downtime does each 'nine' allow?
- Three nines (99.9%) leaves about 43 minutes a month, four nines (99.99%) about 4 minutes, and five nines (99.999%) about 26 seconds. Two nines (99%) leaves about 7 hours. Each extra nine cuts the allowed downtime by roughly ten. See the SLA tier pages for the full per-tier ladder and error budget.
- Is more nines always better?
- No. Past the point users notice, extra nines cost far more than they return, and chasing them can slow delivery enough to hurt the product more than the rare outage would. Set the target to what users need and spend the rest of the effort on shipping.
Want this applied to your stack, not just defined?
The free tools run the numbers; an audit tells you where the real cost and risk are. Book a call, or leave your email and I'll reach out.
Prefer proof first? See how this plays out in real case studies →