The StatusCat blog
Uptime percentages: how much downtime is 99.9%?
What 99.9%, 99.95% and 99.99% uptime actually mean in minutes and hours of allowed downtime — per day, month and year — plus how to pick the right target for your service.
"We offer 99.9% uptime" sounds precise, but it's easy to lose track of what those numbers mean in real minutes of downtime. Here's the plain-English breakdown — and how to choose a target that fits your service.
The downtime cheat sheet
This is the amount of allowed downtime for each uptime level:
| Uptime | Per day | Per month | Per year |
|---|---|---|---|
| 99% ("two nines") | 14m 24s | 7h 18m | 3d 15h |
| 99.9% ("three nines") | 1m 26s | 43m 50s | 8h 46m |
| 99.95% | 43s | 21m 55s | 4h 23m |
| 99.99% ("four nines") | 8.6s | 4m 23s | 52m 36s |
| 99.999% ("five nines") | 0.86s | 26s | 5m 15s |
A useful rule of thumb: each extra nine cuts your allowed downtime by roughly 10× — and tends to increase the cost and engineering effort by more than that.
How uptime is calculated
Uptime % = (total time − downtime) ÷ total time × 100
If your service was down for 22 minutes in a 30-day month:
- Total minutes in 30 days = 43,200
- Uptime = (43,200 − 22) / 43,200 × 100 ≈ 99.949%
So 22 minutes of downtime in a month lands you just under 99.95%.
Which target should you aim for?
- 99.9% — the standard for most SaaS. Achievable with good practices, sensible redundancy, and fast alerting.
- 99.95% — common for paid business tools and light SLAs. Needs redundancy and a real on-call process.
- 99.99% — for revenue-critical or contractual SLAs. Requires multi-region redundancy, automated failover, and mature incident response. Expensive.
- 99.999% — rarely necessary outside infrastructure providers; the cost is enormous.
Don't chase nines for their own sake. Match the target to what your users and contracts actually require, then measure it honestly.
Measuring it in practice
You can't claim an uptime number you don't track. To measure real uptime you need:
- Frequent checks (every minute or 30 seconds) so short outages are captured.
- Re-checks before alerting, so a single failed request doesn't distort your numbers.
- Uptime history and reports to show the actual percentage over time.
StatusCat tracks all of this and can generate SLA reports from it. If you're just getting started, read what uptime monitoring is, then set up your first checks — free for 50 monitors.