Core Reliability
Uptime & SLA Calculator
Turn an uptime target into allowed downtime per day, week, month, and year — or measured downtime into uptime and error budget.
Downtime = (1 − SLA) × periodResult
- Per day
- 1 m 26 s
- Per week
- 10 m 5 s
- Per month (avg.)
- 43 m 48 s
- Per quarter (avg.)
- 2 h 11 m 24 s
- Per year (365 d)
- 8 h 45 m 36 s
- Nines
- 3.0
Downtime = (100% − 99.9%) × 8,760 h = 8 h 45 m 36 s per year
365-day year; month and quarter are average lengths (730 h and 2,190 h).
Uptime and service level agreements
Uptime is the share of a period during which a service or machine was working. A service level agreement (SLA) usually states it as a percentage — 99.9% per calendar month, for example — and the useful way to read that number is as the downtime it leaves you: the downtime budget.
Uptime is the measured form of operational availability. The availability calculator works from MTBF and MTTR; this one works from the percentage and the clock, which is how SLAs and service level objectives (SLOs) are written.
Uptime formulas
Downtime = (1 − uptime target) × period lengthUptime = 1 − downtime ÷ period lengthBudget used = measured downtime ÷ allowed downtimeNines = −log₁₀(1 − uptime)Period lengths here use a 365-day year (8,760 h). Month and quarter are the average lengths, 730 h and 2,190 h. A contract that measures each calendar month allows slightly different downtime in each — at 99.9%, about 40 m 19 s in a 28-day month and 44 m 38 s in a 31-day month.
Allowed downtime by uptime target
| Uptime | Name | Per day | Per week | Per month | Per year |
|---|---|---|---|---|---|
| 99% | Two nines | 14 m 24 s | 1 h 40 m 48 s | 7 h 18 m | 3 d 15 h 36 m |
| 99.5% | 7 m 12 s | 50 m 24 s | 3 h 39 m | 1 d 19 h 48 m | |
| 99.9% | Three nines | 1 m 26 s | 10 m 5 s | 43 m 48 s | 8 h 45 m 36 s |
| 99.95% | 43.2 s | 5 m 2 s | 21 m 54 s | 4 h 22 m 48 s | |
| 99.99% | Four nines | 8.64 s | 1 m | 4 m 23 s | 52 m 34 s |
| 99.999% | Five nines | 864 ms | 6.05 s | 26.3 s | 5 m 15 s |
Each extra nine cuts the allowed downtime by a factor of ten. Going from 99.9% to 99.99% takes a service from roughly three-quarters of an hour of downtime a month to about four minutes — short enough that a single slow manual recovery uses the whole month’s budget.
Worked examples
Example 1 — what does 99.95% allow? Per average month:
(1 − 0.9995) × 730 h = 0.365 h = 21 m 54 sExample 2 — did we meet 99.9% this month? A service was down for 50 minutes in an average month (730 h = 43,800 minutes):
1 − 50 ÷ 43,800 = 99.886%50 min ÷ 43.8 min = 114% — target missed by 6.2 minutesDependencies: series and redundancy
A service that needs several components to all be up can be no more available than their product. Two dependencies at 99.9% each, both required, give at most:
0.999 × 0.999 = 99.80%Two independent redundant copies at 99.9%, where either one is enough:
1 − (1 − 0.999)² = 99.9999%Redundancy only helps as much as the failures are independent — shared power, network, configuration, or deploys put both copies down together. For larger block diagrams use the system reliability calculator.
What to check in an SLA
| Term | Why it matters |
|---|---|
| Measurement window | Monthly, quarterly, or yearly. A yearly 99.9% lets one 8-hour outage pass; a monthly one does not. |
| Exclusions | Planned maintenance windows and force majeure are often excluded from downtime. |
| Definition of “down” | Total outage only, or also degraded performance, error-rate thresholds, or single-region failures. |
| Granularity | Downtime counted per minute or per 5-minute interval rounds short blips differently. |
| Remedy | Usually service credits, not compensation for the business impact of the outage. |