Cloud Server SLA Guide: How 99.9% and 99.99% Uptime Are Counted
An uptime SLA is a downtime budget. 99.9% allows about 43.8 minutes of downtime in an average month, 99.99% about 4.4 minutes and 99% over seven hours. Before comparing hosts, read how uptime is measured, which outages are excluded and how service credits are claimed, because those terms decide the real promise.
For buyers comparing hosting and cloud SLAs who want to turn percentages into minutes and read the fine print.
knowledge
- Read time
- 2 min
- Sections
- 4
- Updated
- 14 Nov 2026
- Written by
- WebJi® team
Last editorial update: 14 November 2026.
How much downtime does each SLA percentage allow?
An uptime percentage is a downtime budget: 99.9% allows about 43.8 minutes of downtime in an average month, 99.99% about 4.4 minutes, and 99% more than seven hours. Convert every promise into minutes before comparing providers.
| SLA | Per month | Per year |
|---|---|---|
| 99% | About 7.3 hours | About 3.65 days |
| 99.5% | About 3.65 hours | About 1.83 days |
| 99.9% | About 43.8 minutes | About 8.76 hours |
| 99.95% | About 21.9 minutes | About 4.38 hours |
| 99.99% | About 4.4 minutes | About 52.6 minutes |
| 99.999% | About 26 seconds | About 5.3 minutes |
The arithmetic is simple: minutes in the period × (100% − SLA%). A 30-day month has 43,200 minutes, so 0.1% of it is 43.2 minutes; providers using a calendar month will show slightly different figures for 28- and 31-day months.
How is uptime actually measured?
Uptime is measured by monitors that check the service at fixed intervals from outside, and the result depends on what is checked, how often and from where. Two monitors can report different uptime for the same month.
- What is checked: a server ping, a TCP port or a real page load. A ping can succeed while the website returns errors.
- Interval: a five-minute check can miss short outages that a one-minute check records.
- Location: checks from one city may report an outage that is really a local network problem.
- Confirmation: good monitors confirm a failure from a second location before counting it.
Run your own independent monitor so you hold evidence if you need to claim; uptime monitoring explains the setup.
Which SLA exclusions change the real numbers?
Exclusions remove certain downtime from the calculation, so the measurement section matters as much as the headline percentage. A 99.9% SLA with broad exclusions can allow more real downtime than a plainly worded 99.5%.
| Exclusion | What to ask the provider |
|---|---|
| Scheduled maintenance | How much notice, how long, how often, at what hour (IST)? |
| Upstream or third-party networks | Where does the provider’s responsibility end? |
| Customer-caused issues | Are plugin errors, overloads and misconfigurations excluded? |
| DDoS and force majeure | Is mitigation included, and what counts as force majeure? |
| Partial degradation | Does a slow but responding service count as down? |
How do SLA service credits work?
A service credit is a discount on a future bill when the provider misses its target, usually calculated as a share of that month’s fee for the affected service and capped. Credits rarely cover the business loss of an outage, so treat them as an incentive, not insurance.
- Read how credits are scaled: many SLAs give larger credits the further the month falls below target.
- Check the claim window — often a set number of days after the month ends.
- Check the cap, which is usually no more than that month’s fee.
- Keep your own monitoring data and incident times in IST to support a claim.
Credits are almost always applied only on request, never automatically, so diarise the review at month end.
Why is your application’s uptime lower than the server SLA?
When a request depends on several components in series — server, database, DNS, payment gateway — their availabilities multiply, so the whole is less available than any single part. Redundancy works the other way and raises availability.
Components in series: 0.999 × 0.999 × 0.999 = 0.997 → about 99.7%
Two servers in parallel: 1 − (0.001 × 0.001) = 0.999999 → about 99.9999%A site on a 99.9% server that also depends on third-party DNS and an external API at 99.9% each can expect roughly 99.7% overall — about 2.2 hours a month. To go higher, add redundancy at the weakest link rather than buying a higher figure on one component. If your site goes down, what to do when your website is down lists the first checks.
What does the WebJi® hosting SLA commit to?
The WebJi® hosting SLA sets a 99.9% monthly uptime target for shared, reseller, VPS and cloud services, with service credits on eligible services when the target is missed. That target allows roughly 43 minutes of unplanned downtime a month, or about 8.76 hours a year.
- Exclusions: scheduled maintenance announced in advance, client-side code, third-party DNS and upstream ISP failures.
- Credits: no percentage is printed; claims go to support@webji.services with the service name and the interval you believe missed target.
- Response: tiered by severity; exact targets are in the client area and order confirmation.
- Evidence: cluster uptime is on a public status page with 90-day history, and independent customer monitors are compared in a dispute.
The full terms are on the WebJi® SLA page. For workloads that need more resilience than one server, cloud hosting plans can be designed with redundancy on request.
Frequently asked questions
Is a 99.99% SLA worth paying extra for?
Only if each minute of downtime costs more than the price difference, as for payment processing or trading systems. For a brochure site or blog, 99.9% is normally enough. Remember that your own application and third-party services also limit real uptime.
Can any host promise 100% uptime?
No real infrastructure runs with zero downtime: hardware fails, software needs patching and networks have incidents. A “100%” promise usually means credits are paid when it is missed. A realistic, measurable target with clear exclusions is easier to hold a provider to.
Does the SLA cover slow performance?
Usually not. Most hosting SLAs define downtime as the service not responding, not as being slow. If speed matters contractually, ask whether partial degradation or response-time thresholds are covered, and in what terms.
How do I calculate uptime for a month myself?
Divide minutes the service was up by total minutes in the month, then multiply by 100. For example, 60 minutes down in a 30-day month (43,200 minutes) gives (43,140 ÷ 43,200) × 100 ≈ 99.86%, which misses a 99.9% target.
SLA terms from the WebJi® SLA page, checked 15 November 2026 · Reviewed by the WebJi® team, Jaipur.
Ready to get started with WebJi®?
Join 12,000+ businesses across India and overseas. Free migration* (subject to terms & approval), INR pricing and 24/7 email support in Hindi & English — average first reply under 60 minutes.