BYTETOOLS

SLA Uptime Calculator

Turn an uptime target into allowed downtime per day, week, month, quarter and year, check achieved uptime, and model composite and redundant availability.

8h 45m 58s
Allowed downtime per year
WindowLengthAllowed downtime at 99.900%In minutes
Per day1d1m 26s1.44
Per week7d10m 5s10.08
Per month30d 10h 30m43m 50s43.83
Per quarter91d 7h 30m2h 11m 29s131.49
Per year365d 6h8h 45m 58s525.96
99.8814%
Achieved uptime
8m 10s
Error budget overrun

52m of downtime in this window is above the 43m 50s the 99.900% target allows — the error budget is 119% spent.

Composite availability — components in series, each optionally duplicated in parallel
100.00000%
100.00000%
99.95000%
99.99000%

Columns: component name · availability % of one unit · number of redundant copies · the resulting availability of that group.

99.94000%
System availability
5h 15m 33s
System downtime per year
Primary database
Weakest link
Yes
Meets the target?
ComponentSingle unitCopies in parallelGroup availabilityDowntime per year
Load balancer99.990%2100.00000%0s
Application tier99.900%3100.00000%0s
Primary database99.950%199.95000%4h 22m 59s
Object storage99.990%199.99000%52m 36s
Whole system (in series)99.94000%5h 15m 33s
Service-credit ladder
If uptime falls below
credit this % of the fee
If uptime falls below
credit this % of the fee
If uptime falls below
credit this % of the fee
10%
Service credit earned
$240.00
Credit value

allowed downtime = window × (1 − target)  ·  achieved = 1 − downtime ÷ window  ·  series = A₁ × A₂ × …  ·  parallel = 1 − (1 − A)ⁿ

Check the numbers against the standard nines table: at 99.9% on a 365.25-day year this calculator gives 8h 45m 36s per year, 43.83 minutes per month, 10.08 minutes per week and 1.44 minutes per day; at 99.99% it gives 52.60 minutes per year and 4.38 per month; at 99.999%, 5.26 minutes per year. Two components at 99% in series come to 98.01%, and the same component duplicated in parallel reaches 99.99%.

All of this runs in your browser; no availability data is uploaded. The year length matters more than people expect — a “month” in an SLA is usually one twelfth of a year rather than a specific calendar month, which is why 99.9% is quoted as both 43.2 and 43.83 minutes depending on the convention. Real contracts also define what counts as downtime (scheduled maintenance is normally excluded), how it is measured, and whether credits must be claimed within a window; read the agreement rather than relying on the arithmetic alone.

What is the SLA Uptime Calculator?

Three nines sounds precise until someone asks how many minutes that is. This calculator converts any uptime target into the allowed downtime for a day, week, month, quarter and year, and works the other way too: give it the minutes you actually lost and it tells you the uptime you achieved and how much of your error budget is gone.

  • Allowed downtime per day, week, month, quarter and year for any target
  • Reverse calculation from observed downtime minutes to achieved uptime
  • Error budget remaining, or the overrun when the target is missed
  • Composite availability in series with redundancy in parallel
  • Weakest-link identification across your component list
  • Editable service-credit ladder applied to a fee for the period

How to use the SLA Uptime Calculator

  1. 1

    Pick an uptime target from the presets or type your own, and choose the year length convention.

  2. 2

    Read the allowed downtime table for a day, week, month, quarter and year.

  3. 3

    Enter observed downtime minutes and the window they cover to get achieved uptime and error budget.

  4. 4

    List your components with their availability and number of redundant copies to get system availability.

  5. 5

    Set the service-credit ladder thresholds and the fee for the period to see the credit earned.

About the SLA Uptime Calculator

Three nines sounds precise until someone asks how many minutes that is. This calculator converts any uptime target into the allowed downtime for a day, week, month, quarter and year, and works the other way too: give it the minutes you actually lost and it tells you the uptime you achieved and how much of your error budget is gone.

It also does the architecture maths. List the components of a service with their individual availability and how many redundant copies each has, and the tool combines them — parallel copies with one minus (one minus A) to the power n, then the resulting groups multiplied together in series. That is how you discover that four solid components in a chain are less reliable than any one of them.

Finally there is an editable service-credit ladder: set the thresholds and credit percentages from your contract and a fee for the period, and see what a bad month actually costs you. Everything runs in your browser and no availability data is uploaded.

Frequently asked questions

How much downtime does 99.9% uptime allow?

About 8 hours 45 minutes a year, or 43.8 minutes a month and 1.44 minutes a day. Four nines cuts that to 52.6 minutes a year and roughly 4.4 minutes a month; five nines allows just 5.26 minutes a year, which is under a second a day.

How do you calculate availability for services in series?

Multiply them. Two components at 99% each give 0.99 times 0.99 = 98.01%, not 99%. Every dependency you add makes the chain worse, which is why the total availability of a system is almost always below the availability of its best component.

How much does redundancy improve availability?

For n independent copies the group availability is one minus (one minus A) to the power n. Two 99% components in parallel reach 99.99% and three reach 99.9999%. The catch is the word independent — shared power, shared network or shared deploys make the real figure much worse than the formula suggests.

Why does the allowed downtime per month vary between calculators?

Because a month is ambiguous. Most SLAs define it as one twelfth of a year, which on a 365.25-day year makes 99.9% equal 43.83 minutes; using a flat 30-day month gives 43.2 instead. The year-length setting here lets you match whichever convention your contract uses.

How do SLA service credits work?

A contract sets thresholds — for example a 10% credit below 99.9% and 25% below 99%. The credit is applied to the fee for the period, and this tool picks the largest tier your achieved uptime fell below. Real contracts usually also require you to claim the credit within a window, so read the agreement.

Related tools