SLA Availability & Uptime Calculation Reference
A Service Level Agreement (SLA) is a formal legal agreement between a service provider and its customers defining expected availability, responsiveness, and financial remedies (e.g. service credits) for outages.
High Availability Uptime & Downtime Table (The "Nines")
The table below outlines the maximum allowable downtime corresponding to standard availability percentages:
| Availability ("Nines") | Downtime per Day | Downtime per Month (30d) | Downtime per Year (365d) |
|---|---|---|---|
| 99.0% (Two Nines) | 14m 24s | 7h 12m 00s | 3d 15h 39m 36s |
| 99.5% | 7m 12s | 3h 36m 00s | 1d 19h 49m 48s |
| 99.9% (Three Nines) | 1m 26s | 43m 12s | 8h 45m 57s |
| 99.95% | 43.2s | 21m 36s | 4h 22m 58s |
| 99.99% (Four Nines) | 8.6s | 4m 19s | 52m 35s |
| 99.999% (Five Nines) | 0.86s | 25.9s | 5m 15s |
Availability Formula
\[\text{Availability} = \frac{\text{Total Time} - \text{Total Downtime}}{\text{Total Time}} \times 100\%\]
Example Calculation:
For a 30-day month (\(\text{Total Time} = 30 \times 24 \times 60 = 43,200 \text{ minutes}\)): - If downtime is \(20\text{ minutes}\): $\(\text{Availability} = \frac{43,200 - 20}{43,200} \times 100 = 99.9537\%\)$
Comparing SLA, SLO, and SLI
graph LR
SLI["SLI (What we measure)<br/>e.g. 99.93% successful HTTP calls"] --> SLO["SLO (Our internal target)<br/>e.g. 99.9% availability"]
SLO --> SLA["SLA (Contractual guarantee)<br/>e.g. 99.5% uptime before penalties apply"]- SLI (Indicator): The raw quantitative measurement of reality.
- SLO (Objective): The internal engineering target designed to protect customer satisfaction.
- SLA (Agreement): The legal/contractual threshold with financial consequences. Always set your internal SLO higher than your external SLA to provide a safety margin.