Free tool, no signup
99.9% uptime calculator
99.9% uptime is sometimes called “three nines.” Here is exactly how much downtime that target allows per day, week, month, and year, plus a calculator below to check any other target.
99.9% uptime allows
- Per day
- 1m 26s
- Per week
- 10m 4s
- Per month
- 43m 47s
- Per year
- 8h 45m
“Per month” uses a 365/12-day average month, so twelve months of downtime always add up to one year of downtime. A 28, 30, or 31-day calendar month allows a slightly different figure.
Error budget and burn rate
Running this as an SLO? Enter the window and the downtime so far to see how much error budget is left, how fast it is burning, and the burn-rate alert thresholds for 99.9%.
53.704% of the error budget left
- Budget for 30 days
- 43m 12s
- Remaining
- 23m 12s
- Burn rate
- 1.39x
- Budget gone in
- 11.6 days
Measured so far: 99.861% available. A burn rate of 1x spends exactly the budget by the end of the window; above 1x it runs out early, and at this pace it runs out before the window ends.
| Alert | Budget spent | Long window | Short window | Burn rate | Error rate over |
|---|---|---|---|---|---|
| Page | 2% | 1 hour | 5 minutes | 14.4x | 1.44% |
| Page | 5% | 6 hours | 30 minutes | 6x | 0.6% |
| Ticket | 10% | 3 days | 6 hours | 1x | 0.1% |
How this is computed
99.9% uptime means the remaining 0.1% of the period is the downtime budget the target allows. We apply that fraction to a day (24 hours), a week (7 days), an average month (365 ÷ 12 days, so twelve months of downtime always add up to one year of downtime instead of drifting with each calendar month’s length), and a year (365 days). No rounding tricks: the table above and the figures on this page come from the exact same formula, computed live.
What this number does not tell you
This is the downtime a 99.9% target permits on paper, not a promise that any particular service is meeting it. Whether a vendor actually stayed inside that budget this month is a measurement question, not an arithmetic one, and it depends on who is doing the measuring and from where. If you have a real SLA with a vendor and want to know what a shortfall would cost, see the downtime cost calculator.
Error budgets and burn-rate alerts, briefly
An SLO’s error budget is the downtime its target allows over a rolling window, usually 28 or 30 days. Burn rate compares how fast you are spending it with an even pace: at 1x the budget lasts exactly the window, at 14.4x a 30-day budget is gone in about two days. Alerting on burn rate rather than on every failure pages people only when the SLO is actually at risk. The thresholds in the table follow the multiwindow policy in Google’s SRE workbook: page when 2% of the budget goes in an hour or 5% in six hours, open a ticket at 10% in three days, and require a short window one twelfth as long to agree so an alert stops soon after the problem does.
Want to know if you’re actually hitting 99.9%?
Knowing the budget is one thing; knowing whether you spent it is another. RealUptime Monitor checks your endpoints on a schedule (four regions on Free, eight on Pro, all ten on Business) and reports your real measured uptime against a target like this one, not just an assumption.
Other uptime targets
99% uptime, 99.5% uptime, 99.95% uptime, 99.99% uptime, 99.999% uptime. See the full SLA calculator overview, or all free tools.