SLA uptime calculator

Turn an uptime percentage into the downtime it allows, work backwards from an outage you had, and see what happens to the total when services depend on each other.

Free · Runs in your browser · Updated 26 September 2026

SLA uptime calculator

What do you want to work out?
For example 99.9, 99.95 or 99.99.
Allowed downtime at 99.9% uptime
PeriodAllowed downtime
Day1 min 26 s
Week10 min 5 s
Month (average)43 min 48 s
Month (30 days)43 min 12 s
Quarter2 h 11 min 24 s
Year (365 days)8 h 45 min 36 s

99.9% allows 8 h 45 min 36 s of downtime a year.

Reference

The nines, in minutes.

The downtime each common uptime level allows. A year is 365 days and the month here is 30 days, the usual choice in SLAs.

Allowed downtime for common uptime levels
UptimePer dayPer weekPer month (30 days)Per year
90%2 h 24 min16 h 48 min3 d36 d 12 h
95%1 h 12 min8 h 24 min1 d 12 h18 d 6 h
99%14 min 24 s1 h 40 min 48 s7 h 12 min3 d 15 h 36 min
99.5%7 min 12 s50 min 24 s3 h 36 min1 d 19 h 48 min
99.9%1 min 26 s10 min 5 s43 min 12 s8 h 45 min 36 s
99.95%43.2 s5 min 2 s21 min 36 s4 h 22 min 48 s
99.99%8.6 s1 min4 min 19 s52 min 34 s
99.999%864 ms6 s25.9 s5 min 15 s
99.9999%86.4 ms604.8 ms2.6 s31.5 s

How it is calculated

Downtime is the leftover fraction of the period.

Allowed downtime

allowed = (1 − uptime ÷ 100) × seconds in the period

For 99.9% over a 365-day year: (1 − 0.999) × 31,536,000 seconds = 31,536 seconds, which is 8 hours 45 minutes 36 seconds. Working backwards from an outage, uptime = (1 − downtime ÷ period) × 100.

  • Day is 86,400 seconds and week is 604,800 seconds.
  • The average month is a year divided by 12, which is 30.42 days. The 30-day month is 30 days. The quarter is a year divided by 4.
  • A leap year has one more day; the calculator uses 365 days throughout.

Chains of services

Dependencies lower the total; redundancy raises it.

All must work (series)

total = A × B × C, each as a fraction

Three services at 99.99%, 99.95% and 99.9% together give 0.9999 × 0.9995 × 0.999 = 99.840065%. The chain is worse than its weakest link, because each service can fail on its own.

Any one is enough (parallel)

total = 1 − (1 − A) × (1 − B)

Two independent copies at 99% give 1 − 0.01 × 0.01 = 99.99%. The word that matters is independent: two copies in one data centre, on one power feed or behind one load balancer do not fail independently, so the real figure is lower.

FAQ

Questions people ask first.

How many minutes of downtime does 99.9% uptime allow?
43 minutes 48 seconds in an average month (a year divided by 12), 43 minutes 12 seconds in a 30-day month, and 8 hours 45 minutes 36 seconds in a 365-day year.
Why does my provider quote a different number for the same percentage?
They are measuring over a different period. A 30-day month, a calendar month of 28 to 31 days and an average month of 30.42 days each give a slightly different allowance. The calculator shows both the average month and the 30-day month so you can match your contract.
What is five nines?
99.999% uptime, which allows about 5 minutes 15 seconds of downtime in a 365-day year, or about 0.9 seconds a day.
How do I work out the uptime of several services together?
If every service must work for a request to succeed, multiply their uptimes: 99.9% times 99.9% is 99.8001%. If any one of several redundant copies is enough, multiply their downtimes instead: two 99% services give 1 − 0.01 × 0.01 = 99.99%, assuming they fail independently.
Does uptime cover planned maintenance?
That depends on the agreement. Many SLAs exclude scheduled maintenance from the calculation, and some count only certain kinds of failure. Read how your provider defines downtime, the window it is measured over and what remedy you get, because the percentage alone does not say.

From the team behind the tools

Need to know what your system can honestly promise?

Daniotech builds web and real-time software. If you are writing an SLA or checking what your architecture can deliver, tell us what you are running.