Service Level Agreements often refer to monthly downtime to calculate credits.
In many cases, "scheduled maintenance" is not included within the availability calculation.
| Availability | Downtime per year | Downtime per month | Downtime per week | Downtime per day |
|---|---|---|---|---|
| 90% | 36.5 days | 72 hours | 16.8 hours | 141 minutes |
| 99% | 3.6 days | 7.2 hours | 1.68 hours | 14.4 min |
| 99.9% | 8.8 hours | 43.2 min | 10.1 min | 86 s |
| 99.99% | 52.6 min | 4.3 min | 1.01 min | 8.6 s |
| 99.999% | 5.3 min | 25.9 s | 6.1 s | 0.86 s |
| 99.9999% | 31.5 s | 2.6 s | 0.61 s | 0.086 s |
Calculations and metrics are great. This being said, what's the nines model does not take into account is timing!
It assumes that all time is worth exactly the same amount to the customers or organization that has deployed the highly available critical system.
Would 1 hour of downtime hurt CNN.com the same way if occurring in the evening of the US election date versus at 4am on a Sunday in July?
Posts
