Convert any uptime percentage — "the nines" — into real downtime per day, week, month, and year. Includes a full reference table for the standard tiers from 90% to 99.9999%.
| Uptime % | Downtime / Month | Downtime / Year |
|---|
Enter an uptime percentage to see downtime figures
The uptime calculator converts an uptime percentage — the classic "nines" you see quoted in hosting plans, cloud provider SLAs, and infrastructure marketing — into a concrete, human-readable amount of allowed downtime per day, per week, per month, and per year. Type in any percentage, from a modest 95% to an extremely demanding 99.9999%, and instantly see how much downtime that permits, alongside a full reference table for every standard tier so you can see exactly where your number sits relative to industry-standard benchmarks like "three nines" (99.9%) or "five nines" (99.999%).
The math is a single formula: downtime = total period length in seconds × (1 − uptime% ÷ 100). A day has 86,400 seconds, a week 604,800, a 30-day month 2,592,000, and a year 31,536,000 — multiply any of those by the "unavailable" fraction of your uptime percentage and you get the maximum downtime allowed in that period while still hitting the stated percentage. The calculator runs this for your custom entry live, and also renders it for the nine standard reference tiers (90%, 95%, 99%, 99.5%, 99.9%, 99.95%, 99.99%, 99.999%, 99.9999%) so you can compare instantly.
It's built for DevOps and SRE teams translating an availability target into monitoring thresholds, developers comparing hosting and cloud vendor SLAs before committing, startup founders choosing a realistic reliability target for launch, and anyone who has ever seen a claim like "99.99% uptime guaranteed" and wondered how many minutes that really is per year. It's equally useful for reliability engineers setting incident-response escalation triggers, and for procurement evaluating whether a vendor's promised availability meets business requirements. Note that this page covers the general nines-to-downtime conversion — for the contractual angle, including service credits and penalties, cross-reference the separate SLA Calculator.
Uptime percentages compress downtime exponentially as they climb toward 100% — each additional "nine" cuts allowed downtime by roughly a factor of ten. That means the jump from 99% to 99.9% is a huge engineering and cost commitment (going from days of allowed downtime per year to hours), and the jump from 99.99% to 99.999% ("five nines") pushes allowed yearly downtime down to just minutes. Understanding what a percentage means in real minutes and hours makes it much easier to evaluate whether a vendor's advertised uptime, or your own system's target, is realistic and adequate for your use case.
How this uptime calculator turns a percentage into real seconds, minutes, and hours
The unavailable fraction (1 − uptime% ÷ 100) is the share of each period the system may be down. At 99.95% the fraction is 0.05% (0.0005), so the allowed downtime is simply that fraction of the period's total seconds — the calculator then formats the result into readable units like "4h 22m 48s" for the year.
The calculator uses a 30-day month (2,592,000 seconds) and a 365-day year (31,536,000 seconds) as standard approximations. Calendar months vary from 28 to 31 days, so real monthly budgets shift slightly depending on the month being measured.
Subtracting the uptime percentage from 100 gives the share of time the system may be down. At 99.95%, that's 0.05% (0.0005). Multiply that fraction by the period's total seconds to get the exact downtime allowed.
Each additional "nine" cuts the unavailable fraction by a factor of ten, which is why downtime shrinks so dramatically from 99.9% to 99.99% and beyond — a 0.1% difference in percentage is a 10x difference in downtime.
From typing a percentage to reading your allowed downtime across every period
Type any value from 0 to 100 into the "Uptime %" field, with up to four decimal places. The default 99.95 is a good starting point for a customer-facing service.
Press the Calculate button to run the conversion. The results panel appears showing the maximum downtime allowed for your percentage.
The large highlighted card shows downtime per year — the figure most commonly quoted in contracts and reliability discussions, e.g. "4h 22m 48s".
Three result boxes break downtime down per day, per week, and per 30-day month, so you can size alerting thresholds and maintenance windows appropriately.
The "Standard Tiers Reference" table lists the nine classic nines tiers. If your value isn't one of them, it's inserted into the table in sorted order and highlighted so you can compare at a glance.
Use Reset to return to the 99.95 default, or just edit the percentage and calculate again to test different scenarios side by side.
Using the calculator's own default — 99.95% uptime — and the real math from its formula
Suppose a SaaS vendor advertises "99.95% uptime" and you want to know how much downtime that actually permits over a day, a week, a 30-day month, and a full year.
Explanation: 99.95% sits between the standard "three nines" (99.9%) and "four nines" (99.99%) tiers. At about four and a half hours of permitted downtime per year, it's a demanding but achievable target for a customer-facing SaaS or API — roughly half the downtime that plain 99.9% allows (8h 45m 36s per year), but still far more than 99.99%'s 52 minutes. When you enter 99.95 in the calculator, the tier table inserts it between those two rows and highlights it, which is exactly the kind of comparison this page makes quick.
What each standard tier means in allowed downtime and engineering effort
| Availability Tier | Allowed Downtime / Year | What It Generally Means |
|---|---|---|
| 99% (two nines) | 3d 15h 36m | Basic availability — visible outages are routine; fine for internal or non-critical tools |
| 99.9% (three nines) | 8h 45m 36s | Standard business target; a common default for SaaS and public web apps |
| 99.95% | 4h 22m 48s | Upgraded customer-facing target with roughly half of 99.9%'s downtime |
| 99.99% (four nines) | 52m 34s | High reliability — telecom and enterprise grade; costly to engineer |
| 99.999% (five nines) | 5m 15s | Mission-critical — reserved for financial, telecom, and core infrastructure |
| 99.9999% (six nines) | 32s | Extreme carrier-grade target with essentially no tolerated interruption |
If your target sits at three nines or below: you have real headroom for planned maintenance, deploys, and incident recovery. Set up basic uptime monitoring and a simple incident process — you don't need a distributed-SRE program yet.
If you're aiming for four or five nines: the engineering cost grows much faster than the percentage improves. Redundancy across zones, failover automation, comprehensive monitoring, and runbooks all become mandatory — and even then, a single significant incident can consume a large share of the annual budget.
This calculator shows theoretical maximums based purely on the percentage math. Actual availability depends on how uptime is measured (per-request, per-node, region-wide), whether scheduled maintenance is excluded, and your specific SLA terms. For service-credit and penalty math, see the SLA Calculator.
Where a percentage-to-downtime conversion genuinely helps
Translate a vendor's advertised uptime into a concrete per-month downtime budget your monitoring stack can track.
Know exactly how many minutes you can spend in a major incident each month before blowing your availability target.
Give engineering teams a clear, quantified downtime allowance — error budgets — instead of an abstract percentage.
Normalize competing hosting, cloud, and API offers to the same minutes-per-year unit for a fair comparison.
Check how much scheduled downtime your chosen target actually permits before planning deploys and upgrades.
Estimate how a chain of 99.9% services compounds into a much lower end-to-end availability figure.
Justify multi-zone or multi-region redundancy spend when a single-region architecture can't hit four nines.
Turn "99.99% uptime guaranteed" into a defensible expectation before signing a vendor agreement.
Present downtime in minutes and hours rather than abstract percentages to non-technical leadership.
Check whether a real outage exceeded your availability budget and decide whether to escalate investment.
Show teams and students why "the nines" are exponentially harder — not linearly — as they climb.
Pair with the SLA Calculator to move from allowed-downtime math to service credits and financial penalties.
What this uptime calculator does well, and where raw percentage math stops short
Standard nines tiers converted to allowed downtime across every period
| Uptime % | Downtime / Day | Downtime / Month (30d) | Downtime / Year |
|---|---|---|---|
| 99% (two nines) | 14m 24s | 7h 12m | 3d 15h 36m |
| 99.5% | 7m 12s | 3h 36m | 1d 19h 48m |
| 99.9% (three nines) | 1m 26s | 43m 12s | 8h 45m 36s |
| 99.95% | 43s | 21m 36s | 4h 22m 48s |
| 99.99% (four nines) | 9s | 4m 19s | 52m 34s |
| 99.999% (five nines) | 864ms | 26s | 5m 15s |
Summary: This uptime calculator converts any availability percentage — the classic "nines" — into allowed downtime per day, week, month, and year, with a full standard-tier reference table so you can compare your number against the benchmarks. For the contractual side — service credits, penalties, and the cost of missing a target — pair it with the SLA Calculator, and use the API Rate Limit Calculator to check whether your traffic can actually sustain the availability you're promising.
Common questions about uptime percentages and downtime conversions
Official documentation to complement this calculator — always verify against your vendor's actual SLA terms
Explore other developer & tech tools