Skip to main content
Tutorial20 November 2026· 7 min read

How to calculate uptime and downtime percentage

The uptime formula

Uptime percentage is the share of a period in which a service was working. Subtract the downtime from the total time, divide by the total time and multiply by 100:

Uptime % = (total time − downtime) ÷ total time × 100

Use the same unit for both numbers, and minutes are the easiest. A day has 1,440 minutes, so a 30-day month has 43,200. If a website was down for 50 minutes in a 30-day month:

(43,200 − 50) ÷ 43,200 × 100 = 99.88%

Downtime percentage is simply what is left: 100% minus the uptime, or downtime ÷ total time × 100. In the example, 50 ÷ 43,200 × 100 is 0.12%.

How much downtime an uptime target allows

To turn a target into time, work the formula backwards: allowed downtime = (100 − uptime %) ÷ 100 × total time. For 99.9% in a 30-day month that is 0.001 × 43,200 = 43.2 minutes, or 43 minutes 12 seconds.

The answer depends on how long the period is, so here is each common target for the three month lengths and a year:

Uptime28-day month30-day month31-day month365-day year
99%6h 43m 12s7h 12m7h 26m 24s3 days 15h 36m
99.5%3h 21m 36s3h 36m3h 43m 12s1 day 19h 48m
99.9%40m 19s43m 12s44m 38s8h 45m 36s
99.95%20m 10s21m 36s22m 19s4h 22m 48s
99.99%4m 2s4m 19s4m 28s52m 34s
99.999%24s26s27s5m 15s

Seconds are rounded to the nearest whole second. Each extra nine cuts the allowance by ten: going from 99.9% to 99.99% takes a 30-day month from 43 minutes to under 5. To try your own target, or to put a revenue figure on the downtime, use our uptime and downtime calculator.

Why calculators give different answers

If one source says 99.9% allows 43.2 minutes a month and another says 43.8, neither is wrong. They assume different months:

  • A 30-day month has 43,200 minutes, so 0.1% is 43.2 minutes (43m 12s).
  • An average month is a year divided by 12: 525,600 ÷ 12 = 43,800 minutes, so 0.1% is 43.8 minutes (43m 48s).
  • A 31-day month has 44,640 minutes, so 0.1% is 44.6 minutes (44m 38s).

A year works the same way: 365 days give 8 hours 45 minutes 36 seconds at 99.9%, and a 366-day leap year gives 8 hours 47 minutes 2 seconds. Our uptime calculator averages over a 365.25-day year to allow for leap years, so its monthly figure is a twelfth of that. For a report or an agreement, state the period you measure over, ideally the actual calendar month, so everyone gets the same number.

Calculating uptime from monitoring checks

A monitoring service does not watch a site continuously. It checks at a fixed interval and records whether each check passed, so its uptime figure is usually the share of checks that passed:

Uptime % = checks that passed ÷ all checks × 100

That makes the interval matter. With one check a minute, a 30-day month has 43,200 checks and one failed check moves the figure by about 0.0023%. With one check every five minutes there are 8,640 checks, and one failed check is worth about 0.0116%. A longer interval also means coarser measurement: an outage shorter than the interval can be missed entirely, and one that is caught may be counted as a whole interval of downtime.

PulseStack™'s 30-day uptime figure works this way: it is the share of a monitor's checks in the last 30 days that found it up, leaving out any checks made during a maintenance window you have set. Checks run every 5 minutes on the Free plan, every 60 seconds on Starter and every 30 seconds on Pro and above. See website monitoring for how the checks work, and pricing for the plans.

Decide what counts as downtime first

The arithmetic is the easy part. The figure only means something if you have decided what counts as down, and apply the same rules every time:

  • Planned maintenance. Some service level agreements leave out maintenance announced in advance; others count everything. Use whatever your agreement says, and keep internal reports on the same basis.
  • Partial outages. Is a site with a broken checkout but a working home page up or down? Decide which pages or functions count, and monitor those.
  • Slow responses. A page that takes 30 seconds to load counts as up by the formula but is of little use to a visitor. Some teams count responses slower than a threshold as downtime.
  • Unknown time. If the monitoring itself was unavailable, the period was neither up nor down. Say whether you leave it out or count it.
  • Where it was measured from. A site can fail in one region and work in another. Checks from more than one location, confirmed before an alert, avoid counting one network's fault as a full outage.

For how uptime targets become internal objectives and contractual promises, read SLA, SLO and SLI explained.

Uptime formulas for Excel and Google Sheets

The same formulas work in Microsoft Excel and Google Sheets. Put one month per row:

CellWhat goes in itFormula
D2Any date in the month, for example 01/11/2026typed in
A2Total minutes in that month=DAY(EOMONTH(D2,0))*1440
B2Minutes of downtime in the monthtyped in, or a sum of your incident rows
E2Uptime percentage=1-B2/A2, formatted as a percentage with 3 decimal places
C2Your target, for example 99.9%typed in
F2Downtime the target allows, in minutes=(1-C2)*A2
G2Minutes left before the target is missed=F2-B2

If you log each incident with a start and end date and time, the length of each one in minutes is =(end-start)*1440, because spreadsheets store dates as days. Add those up for the month's downtime. To show minutes as hours, minutes and seconds, divide by 1,440 and use the custom number format [h]:mm:ss.

A worked example

Suppose a website had three outages in November, which has 30 days: 12 minutes, 7 minutes and 26 minutes. These are example figures to show the method.

  1. Total time: 30 × 1,440 = 43,200 minutes.
  2. Downtime: 12 + 7 + 26 = 45 minutes.
  3. Uptime: (43,200 − 45) ÷ 43,200 × 100 = 99.896%, which rounds to 99.90% at two decimal places.
  4. Against a 99.9% target the month allowed 43.2 minutes, so the site missed it by 1.8 minutes.

The example shows why rounding matters: at two decimal places the report says 99.90% and looks like a pass. Compare minutes against the allowance, or keep three decimal places, before you call a month a success. To work out what those 45 minutes cost, see how to calculate the true cost of downtime.

Sources

The figures on this page are arithmetic, with the working shown. The spreadsheet functions are documented by Microsoft and Google, checked on 20 November 2026:

Frequently asked questions

How do you calculate uptime percentage?

Take the total time in the period, subtract the time the service was down, divide by the total time and multiply by 100. A 30-day month has 43,200 minutes, so 50 minutes of downtime gives (43,200 minus 50) divided by 43,200, which is 99.88% uptime.

How much downtime is 99.9% uptime?

0.1% of the period: 43 minutes 12 seconds in a 30-day month, 44 minutes 38 seconds in a 31-day month and 8 hours 45 minutes 36 seconds in a 365-day year.

How do I calculate uptime in Excel or Google Sheets?

Put the total minutes in one cell and the minutes of downtime in another, then use =1-B2/A2 and format the cell as a percentage. The same formula works in both. To get a month's total minutes from any date in it, use =DAY(EOMONTH(D2,0))*1440.

Does planned maintenance count as downtime?

It depends on how the figure is defined. Some service level agreements leave out maintenance that was announced in advance, but the agreement's own wording decides. Whatever you choose, apply it the same way every month and say which rule you used.

Why do uptime calculators give different downtime figures?

Mostly because they assume different month lengths. 99.9% of a 30-day month is 43.2 minutes, of an average month (365 days divided by 12) 43.8 minutes, and of a 31-day month 44.6 minutes. All three are right for their assumption. Our own calculator averages over a 365.25-day year, so its month is a twelfth of that, about 43.8 minutes at 99.9%.

Share this page

Start monitoring your infrastructure today

5 free monitors, no credit card needed. Set up in under 30 seconds.

Get started free