An error budget
The allowed downtime is a budget. Releases, migrations and incidents all spend it, so a higher target means slower, more careful changes.
Free tool
Type an availability target and see how much downtime it allows. Add the services you depend on to see what you can actually promise. It all runs in your browser.
| Target | Per day | Per week | Per month | Per year |
|---|---|---|---|---|
| 99% | 14 min 24 s | 1 h 40 min 48 s | 7 h 18 min 18 s | 3 d 15 h 39 min |
| 99.5% | 7 min 12 s | 50 min 24 s | 3 h 39 min 9 s | 1 d 19 h 49 min |
| 99.9% | 1 min 26 s | 10 min 5 s | 43 min 50 s | 8 h 45 min 58 s |
| 99.95% | 43 s | 5 min 2 s | 21 min 55 s | 4 h 22 min 59 s |
| 99.99% | 8.6 s | 1 min | 4 min 23 s | 52 min 36 s |
| 99.999% | 864 ms | 6 s | 26 s | 5 min 16 s |
A year is counted as 365.25 days, a month as a twelfth of that and a quarter as a fourth.
The allowed downtime is a budget. Releases, migrations and incidents all spend it, so a higher target means slower, more careful changes.
Going from 99.9% to 99.99% cuts allowed downtime tenfold. That usually takes redundancy, automatic failover and on-call, not just better code.
An SLA should sit below what you actually achieve, and below what your providers promise you, with room for the dependencies you add.
We design deployments, monitoring and failover around the availability you need, and no more than you need.