Free uptime calculator: downtime allowed by a service level
Enter an uptime target or a downtime. The calculator shows the downtime it allows in each period, the uptime you reached, or the uptime of components that depend on each other.
Free, no account needed. Updated .
Short answer
Allowed downtime is (100% − uptime %) × the length of the period. At 99.5% uptime a service can be down for about 3 hours 39 minutes per month or 1 day 19 hours per year. When a service needs several components to work, multiply their uptimes to get the uptime of the whole.
How to use the uptime calculator
- Pick a calculation. Uptime to downtime shows how much downtime a target allows. Downtime to uptime turns an outage into the uptime you reached. Combined uptime shows the uptime of a service that needs several components to work.
- Enter the numbers. Type a percentage, or the hours, minutes and seconds of downtime and the period. The result updates as you type.
- Compare with your agreement. Check how your service level agreement measures uptime before you compare: the period, what counts as down and what is excluded.
The calculator runs in your browser. Your numbers are never sent to a server.
The uptime formula
Allowed downtime = (100% − uptime %) × length of the period. The reverse is uptime % = (1 − downtime ÷ length of the period) × 100.
The calculator uses a year of 365.25 days, so leap years count. A month is one twelfth of that year (about 30.44 days) and a quarter is one fourth (about 91.31 days). Many agreements use calendar months instead. A 31-day month allows a little more downtime than a 28-day month at the same percentage, so check which one yours uses.
Allowed downtime for common targets
| Uptime | Per month | Per year |
|---|---|---|
| 95% | 1 d 12 h 31 min 30 s | 18 d 6 h 18 min |
| 99% | 7 h 18 min 18 s | 3 d 15 h 39 min 36 s |
| 99.5% | 3 h 39 min 9 s | 1 d 19 h 49 min 48 s |
| 99.8% | 1 h 27 min 40 s | 17 h 31 min 55 s |
Each extra nine after the decimal point cuts the allowed downtime by a factor of ten. Each step is much harder to reach than the one before, because a single outage of a few hours can use the whole budget for the year.
A worked example
One target. At 99.5% uptime, 0.5% of a day is 432 seconds, or 7 minutes 12 seconds. Over a week it is 50 minutes 24 seconds, over a month 3 hours 39 minutes 9 seconds, over a quarter 10 hours 57 minutes 27 seconds and over a year 1 day 19 hours 49 minutes 48 seconds.
One outage. A service is down for 2 hours in a month. That is 7,200 seconds of a 2,629,800-second month, so uptime for the month is 99.7262%.
Components in series. A web app needs its hosting (99.8%), its database (99.8%) and a payment provider (99.5%) to work. The combined uptime is 0.998 × 0.998 × 0.995 = 99.1024%. That allows about 6 hours 33 minutes of downtime per month, or about 3 days 6 hours per year. The combined figure is lower than every single component, and lower than any of their targets.
Series and redundancy
When a service needs all its parts, multiply their uptimes. This is why a chain of dependencies needs each part to be better than the target for the whole. To promise 99.5% for the whole service, the parts together must reach it, and every new dependency lowers the result.
Redundancy works the other way. When any one of several copies is enough, the combined uptime is 1 minus the product of their downtime shares. Two independent copies at 90% each give 1 − 0.1 × 0.1 = 99%. The formula assumes the copies fail independently. Copies that share a network, a region or a release fail together, so in practice they gain less.
What to put in a service level agreement
A service level agreement turns the percentage into a promise that both sides read the same way. A useful template covers:
- Scope. Which services and features the target covers, and which it does not.
- Measurement. How uptime is measured: from where, how often, and what counts as down (no answer, an error, or a response slower than a stated limit).
- Period. Calendar month or rolling 30 days, and the time zone.
- Exclusions. Maintenance announced in advance, outages caused by the customer, and events outside your control.
- Target and remedy. The uptime percentage, and the credit or other remedy when it is missed.
- Support response. How customers report an outage, and how fast you reply by severity.
- Reporting. Where customers see the measured uptime, such as a status page or a monthly report.
What the calculator does not do
- It does not measure your uptime. It converts numbers. Use a monitoring service that checks from more than one place to measure it.
- It treats all downtime as equal. An hour down at peak time costs more than an hour at night. Some teams also count partial outages or slow pages, which this calculator does not model.
- It assumes components fail independently. Shared causes make real combined uptime lower than the product suggests.
Uptime and your support team
During an outage, contacts rise fast and many ask the same question. A short status page and a help page that says what to do keep many of those contacts away from your queue. To plan the team for those peaks, use the Erlang calculator.
Frequently asked questions
How do you calculate allowed downtime from an uptime percentage?
Subtract the uptime from 100% and multiply by the length of the period. At 99.5% uptime, 0.5% of a 30.44-day month is about 3 hours 39 minutes. The calculator does this for a day, week, month, quarter and year.
How long is a month in the calculation?
A month is one twelfth of a 365.25-day year, about 30.44 days, and a quarter is one fourth. Some contracts measure calendar months instead, so check which one your agreement uses: a 31-day month allows a little more downtime than a 28-day one.
How do I combine the uptime of several services?
If your service needs all of them to work, multiply their uptimes. Three components at 99.8%, 99.8% and 99.5% give about 99.10%. The result is always lower than the weakest component, so a chain of dependencies needs each part to be better than the target.
Does redundancy raise uptime?
Yes, when one copy is enough to serve. The combined uptime is 1 minus the product of the downtime shares: two independent copies at 90% each give 1 − 0.1 × 0.1 = 99%. Copies that share a power supply, a network or a deployment fail together, so they help less than the formula says.
What should a service level agreement say about uptime?
How uptime is measured (from where, how often and what counts as down), the period, what is excluded such as announced maintenance, the target, how customers report an outage, and the credit or remedy when the target is missed.
Is planned maintenance counted as downtime?
Only if your agreement says so. Many agreements exclude maintenance announced in advance. If yours does, subtract the maintenance windows before you compare the measured downtime with the allowed downtime.
Newsletter
Support notes, once a month.
One email with what we learn about help centers, honest AI answers and support teams. Unsubscribe at any time.
Related
- Free toolErlang C Staffing CalculatorEnter your contact volume, average handle time and target. The calculator finds the smallest team that meets it and shows what one agent more or less would change.
- Free toolCSAT & CES CalculatorEnter how many customers gave each rating. The CSAT or effort score, the breakdown and the margin of error update as you type.
- Free toolNPS CalculatorEnter how many people gave each answer, or paste the raw scores. The score, the breakdown and the margin of error update as you type.