SLA & Uptime Calculator: How Much Downtime Is 99.99%?
Enter an uptime SLA percentage and get the exact allowed downtime for day, week, month, and year. Use this to compare 99.9, 99.99, and 99.999 uptime before you promise an SLA.
Allowed downtime
8.6s
Uptime in this window: 23h 59m 51s
Allowed downtime
1m 0s
Uptime in this window: 6d 23h 58m 60s
Allowed downtime
4m 19s
Uptime in this window: 29d 23h 55m 41s
Allowed downtime
52m 34s
Uptime in this window: 364d 23h 7m 26s
Quick SLA comparison
| Uptime | Downtime / year | Downtime / month |
|---|---|---|
| 99% | 3d 15h 36m | 7h 18m |
| 99.9% | 8h 46m | 43m 50s |
| 99.99% | 52m 34s | 4m 23s |
| 99.999% | 5m 15s | 26s |
How to calculate uptime and downtime
Downtime % = 100% - Uptime %
Downtime = Total time x (1 - Uptime % / 100)
Use these totals in seconds for common windows: day = 86400, week = 604800, month = 2592000, year = 31536000.
Worked example: 99.9% uptime for one week
1) Convert uptime percent to fraction: 99.9 / 100 = 0.999.
2) Weekly total seconds: 60 x 60 x 24 x 7 = 604800.
3) Weekly uptime: 0.999 x 604800 = 604195.2 seconds.
4) Weekly downtime: 604800 - 604195.2 = 604.8 seconds (10m 5s).
The nines of availability
"Nines" is shorthand for how many 9s appear in an uptime percentage. Each additional nine cuts allowed annual downtime by roughly 10x.
Two nines (99%) - 3 days 15 hours per year. Suitable for internal tools.
Three nines (99.9%) - 8 hours 46 minutes per year. Common SaaS baseline.
Four nines (99.99%) - 52 minutes per year. Standard for production APIs.
Five nines (99.999%) - 5 minutes per year. Requires active-active redundancy.
Frequently asked questions
What does 99.99% uptime mean in downtime?
What is the uptime and downtime formula?
What is five nines uptime?
Should I target 99.9% or 99.99% uptime?
How to calculate SLA downtime from an uptime percentage
The formula is straightforward: multiply the total time period by the downtime fraction.
Example for 99.9% over one year:
Reference time periods in seconds: 1 day = 86,400 · 1 week = 604,800 · 1 month (30d) = 2,592,000 · 1 year = 31,536,000.
99.9% vs 99.99% vs 99.999% uptime - a full comparison
| SLA | Nines | Per day | Per month | Per year |
|---|---|---|---|---|
| 99% | Two nines | 14m 24s | 7h 18m | 3d 15h 39m |
| 99.5% | - | 7m 12s | 3h 39m | 1d 19h 49m |
| 99.9% | Three nines | 1m 26s | 43m 50s | 8h 45m 57s |
| 99.95% | - | 43s | 21m 55s | 4h 22m 58s |
| 99.99% | Four nines | 8.6s | 4m 23s | 52m 35s |
| 99.999% | Five nines | 0.86s | 26s | 5m 15s |
SLA vs SLO vs SLI - what is the difference?
SLI - Service Level Indicator
A metric you measure: request success rate, response latency, error rate. SLIs are the raw numbers - the inputs to everything else. Examples: "percentage of requests completing under 200ms" or "percentage of HTTP requests returning 2xx."
SLO - Service Level Objective
An internal target you set for an SLI: "99.9% of requests should complete under 200ms over a 28-day window." SLOs are commitments you make to yourselves - they drive engineering decisions and define when an incident becomes a priority.
SLA - Service Level Agreement
A contractual commitment to a customer with financial consequences for violation (credits, refunds). SLAs are typically set lower than SLOs - you promise customers 99.9% while targeting 99.95% internally, so normal operational variance does not trigger payouts.
Rule of thumb
Set your SLA 0.05%–0.1% below your SLO. If you target 99.95% internally, promise customers 99.9%. The gap is your error budget - it absorbs planned maintenance, deploy windows, and minor incidents without triggering SLA violations.
What infrastructure does each uptime tier require?
99.9% (three nines) - 8h 46m/year
Achievable with a single region and basic redundancy. A single server with automatic restarts, a managed database with daily backups, and a CDN in front handles most failure modes. Planned maintenance windows fit within the allowance.
99.99% (four nines) - 52 min/year
Requires active redundancy: multiple availability zones, auto-scaling, health-checked load balancers, automated failover for the database layer, and zero-downtime deployments. A single AZ failure cannot take the service down.
99.999% (five nines) - 5 min/year
Requires multi-region active-active architecture with sub-second failover. No maintenance window is available - every component must be replaceable without downtime. Typically requires dedicated SRE team and significant infrastructure spend. Realistic only for services where downtime causes direct, large-scale financial loss.
What monitoring interval do you need to detect downtime within your SLA?
Your monitoring check interval determines the maximum undetected downtime - the gap between an outage starting and your system detecting it. For tight SLAs, this gap matters.
| SLA | Monthly allowance | Recommended check interval |
|---|---|---|
| 99% | 7h 18m | 5 minutes is fine |
| 99.9% | 43m 50s | 1 minute recommended |
| 99.99% | 4m 23s | 30 seconds or faster |
| 99.999% | 26 seconds | Synthetic probes + real-user monitoring |
Vantaj checks every 30 seconds from multiple regions by default, with an alert sent within 30 seconds of detection - fast enough for 99.99% SLA commitments.