Learn
What does 99.9% uptime mean?
Uptime percentage is the share of a period during which a service was available, calculated as available time divided by total time. 99.9% uptime, often called three nines, means the service was unavailable for no more than 0.1% of the period: roughly 43 minutes in a 30-day month or about 8 hours 46 minutes in a year.
How is uptime percentage calculated?
The formula is simple: uptime percentage equals the time the service was up divided by the total time in the period, multiplied by one hundred. A 30-day month contains 43,200 minutes. If a service was down for 43 minutes and 12 seconds, it was up for 43,156.8 minutes, which is 99.9% of the month.
In practice availability is sampled, not measured continuously. A monitor that checks every five minutes and sees one failed check attributes about five minutes of downtime, even if the real outage lasted ninety seconds, because that check represents its full share of the period. A thirty-second monitor records the same incident far more precisely, which is one reason two tools can report different numbers for the same service.
Allowed downtime at each level
The table below uses a 30-day month (43,200 minutes) and a 365-day year (525,600 minutes). Each extra nine cuts the allowed downtime by a factor of ten, which is why the jump from 99.9% to 99.99% is far harder than the numbers suggest.
| Uptime | Allowed downtime per month | Allowed downtime per year |
|---|---|---|
| 99% | 7 h 12 min | 3 d 15 h 36 min |
| 99.9% | 43 min 12 s | 8 h 45 min 36 s |
| 99.95% | 21 min 36 s | 4 h 22 min 48 s |
| 99.99% | 4 min 19 s | 52 min 34 s |
Why the measurement window matters
The same percentage means different things over different windows. 99.9% over a year permits a single eight-hour outage, as long as the rest of the year is clean. 99.9% measured monthly permits at most 43 minutes in any one month, so that same eight-hour outage breaks the target for the month in which it happened, however good the other eleven months were. Contracts almost always specify the window, and monthly is the common choice.
Daily figures are less forgiving still: at 99.9%, one day allows only about 86 seconds. This is why a 7, 30 or 90-day uptime bar on a status page is more informative than a single headline figure: it shows when the problems happened as well as how much.
Uptime percentage versus an SLA
An uptime percentage is a measurement. A service level agreement (SLA) is a promise, usually with a definition of downtime, a measurement window, a list of exclusions and a remedy such as service credits if the promise is broken. Most SLAs exclude scheduled maintenance that was announced in advance, outages caused by the customer, and events outside the provider's control. When comparing providers, read the exclusions and the window as carefully as the headline number.
A service level objective (SLO) is an internal target, often stricter than the public SLA, that a team uses to decide when reliability work must take priority over features.
Maintenance windows and exclusions
Planned work that takes a service offline is not usually counted as downtime, provided it was scheduled and announced. Monitoring tools handle this with maintenance windows: periods during which checks keep running, so you still see what happened, but alerts are muted and any downtime is excluded from the uptime calculation. Excluding maintenance keeps the figure honest about unplanned failures while recognising that deliberate downtime is a different thing.
The flip side is that a maintenance window must be genuinely planned. Opening one retrospectively to hide an outage defeats the purpose of measuring, and affected customers will notice the mismatch.
How MoniterMySite handles this
MoniterMySite calculates uptime from checks, so the check interval sets the resolution: a five-minute monitor shows a bigger dip for the same incident than a thirty-second one. Maintenance windows defined under Maintenance in the dashboard can be one-off or repeat daily, weekly or monthly, and apply to all monitors or a selection; during a window checks keep running, alerts are muted and downtime is excluded from uptime figures. Status pages show uptime bars for 7, 30 or 90 days alongside the percentage, and the Summit plan includes SLA reports.
Frequently asked questions
Is 99.9% uptime good?
For most websites and SaaS products, yes. It allows roughly 43 minutes of downtime a month, which is enough to absorb a short incident. Platforms where every minute is costly, such as payments or infrastructure APIs, usually aim for 99.95% or 99.99%.
What is the difference between 99.9% and 99.99%?
A factor of ten. 99.9% allows about 43 minutes of downtime per 30-day month; 99.99% allows about 4 minutes 19 seconds. Reaching four nines generally requires redundancy, automated failover and very fast detection.
Does 100% uptime exist?
Over a short window, certainly. Over a year, almost never, and a provider promising it should be read as promising service credits rather than perfection. Every system has some failure modes it cannot avoid.
Why do two monitoring tools report different uptime for the same site?
Different check intervals, different locations, different fail thresholds and different treatment of maintenance all change the result. A tool checking every five minutes from one region will record outages more coarsely than one checking every thirty seconds from three.
Keep reading
See it in practice.
Add your first monitor on the free plan in under a minute. Multi-region confirmation is included from the Launch plan.