Updated 25 October 2026 · By the Upcertain team
The downtime each target allows
| Uptime | Per day | Per week | Per 30-day month | Per year |
|---|---|---|---|---|
| 99% | 14 m 24 s | 1 h 40 m 48 s | 7 h 12 m | 3 d 15 h 36 m |
| 99.5% | 7 m 12 s | 50 m 24 s | 3 h 36 m | 1 d 19 h 48 m |
| 99.9% | 1 m 26 s | 10 m 5 s | 43 m 12 s | 8 h 45 m 36 s |
| 99.95% | 43 s | 5 m 2 s | 21 m 36 s | 4 h 22 m 48 s |
| 99.99% | 8.6 s | 1 m 0.5 s | 4 m 19 s | 52 m 34 s |
| 99.999% | 0.9 s | 6 s | 25.9 s | 5 m 15 s |
The formula: allowed downtime = (1 - target) x length of the period. A year here is 365 days and a month 30 days. Using the average calendar month (30.4 days) gives slightly more: 43 m 50 s at 99.9%. Try other targets and periods in the uptime calculator.
Reading the table
- 99.9% ("three nines") allows one 43-minute outage a month, or a two-minute blip every day. It is a common target for business software and many SLAs.
- 99.95% halves that: about 22 minutes a month.
- 99.99% ("four nines") allows about four minutes a month. Detecting and confirming an outage can use up a good part of that before anyone starts fixing it.
- 99.999% allows about five minutes a year. It needs automatic failover; no person can respond that fast.
Measured uptime depends on how you measure
The same service can report different uptime depending on the monitor:
- Check interval. A monitor checking every 5 minutes misses most 1- and 2-minute outages, so it reports higher uptime than one checking every 30 seconds. See how often to check.
- What counts as down. A slow page, a 200 with an error message on it, or one region failing may or may not count, depending on the check.
- Maintenance. Many SLAs exclude scheduled maintenance. Upcertain counts time inside a maintenance window as maintenance, not downtime.
- The period. 99.9% over a year allows a single eight-hour outage that would break a monthly 99.9% target ten times over.
Setting a target you can keep
- Measure first. Run a monitor for a month and see what you actually achieve, including deploys.
- Set an internal target above what you promise customers, so a bad month is a warning rather than a breach.
- Budget the downtime. At 99.9% you have 43 minutes a month for incidents and risky changes together. If deploys cost a minute each, twenty deploys a month use half the budget.
- Publish it honestly. A status page with 90 days of history shows customers the real number.
Uptime is not the whole story
Two services with 99.9% uptime can feel very different: one has a single 43-minute outage a month, the other a 90-second blip every night. Uptime also says nothing about speed; a page that takes twelve seconds to load is "up". Watch the number of incidents and how long they last as well as the percentage.
Questions
How much downtime is 99.9% per year?
8 hours, 45 minutes and 36 seconds in a 365-day year.
How much downtime is 99.9% per month?
43 minutes 12 seconds in a 30-day month, or about 43 minutes 50 seconds in an average calendar month.
Is 99.9% uptime good?
For most business websites and SaaS products it is a reasonable, common target. Payment systems and infrastructure other companies build on usually aim higher.