Uptime is the share of a measured period in which a service answers requests correctly, written as a percentage such as 99.9%. Over a 30-day month of 43,200 minutes, 99.9% allows about 43 minutes of downtime. An SLA defines how that figure is measured and what you get when it is missed.
Uptime Defined: What Counts as Down
A site is down when a request does not get the right answer: the connection is refused, it times out, DNS fails to resolve, or the server returns a 5xx. A slow page is not downtime under most contracts, and a maintenance window announced in advance is normally excluded from the calculation.
The definition matters because a percentage is only as good as its measurement window. "99.9% monthly" and "99.9% annually" allow very different amounts of failure, which is why an SLA names the window before it names the number.
The Nines: 99%, 99.9% and 99.99% in Downtime Minutes
A 30-day month has 43,200 minutes, so 99.9% allows 43.2 minutes down and 99.99% allows 4.3. The same percentages stretch to hours and days over a full year of 525,600 minutes.
| Uptime | Downtime per 30 days | Downtime per year | In plain words |
|---|---|---|---|
| 99% | 432 minutes (7 h 12 m) | 3.65 days | Consumer-grade |
| 99.9% | 43.2 minutes | 8 h 46 m | The hosting standard, and WebJi's target |
| 99.99% | 4.3 minutes | 53 minutes | Enterprise-grade |
| 99.999% | 26 seconds | 5.3 minutes | Carrier-grade |
The arithmetic is one line: 43,200 x 0.001 = 43.2. That is why the jump from three nines to four is expensive, since it cuts the allowed failure time by a factor of ten while the engineering cost rises by more than that.
What an Uptime SLA Promises and What It Excludes
An SLA states four things: the target, the measurement window, the exclusions, and the remedy. The remedy is usually a service credit on a future invoice, not a refund and never compensation for lost revenue, so read that clause before you compare two hosts on the headline number.
- Planned maintenance windows, often announced 24 to 48 hours ahead.
- Failures upstream of the host: your DNS provider, your ISP, a CDN or the visitor's own network.
- Customer code, a plugin that crashes PHP, or an account suspended for resource overuse.
- Force majeure and, on many contracts, large-scale DDoS attacks.
How Hosts Measure Uptime: External Probes vs Internal Checks
Most published uptime comes from external HTTP probes that request a URL every one to five minutes from one or more regions. Internal checks watch the same server from inside the network, so they can miss a routing or DNS problem that visitors experience.
The sampling interval decides what the number can see. A five-minute probe that fires at :00 and :05 will miss a two-minute outage at :02 entirely, and the month still reports 100%. Ask for the interval and the number of locations before you treat a percentage as a promise about your visitors.
Monitoring Your Own Site: Checks, Alerts and Thresholds
Independent monitoring is free and takes ten minutes. Check the pages that earn money, alert after two or three consecutive failures, and keep the log, because that log is the evidence an SLA claim needs.
- UptimeRobot (free tier) runs HTTP checks every five minutes and alerts by email.
- StatusCake or Better Stack offer similar free tiers with longer retention.
- Check key pages, not just the homepage: the checkout, the enquiry form, the login.
- Set the threshold to two failures, so one blip does not wake you but a real outage does.
For a store taking Rs 1,00,000 a month, one outage hour averages about Rs 140 in lost sales, before the trust cost. Ten minutes of setup turns an outage from a surprise into a notification.
The Cost of Downtime: Revenue, SEO and Trust
Revenue loss is the visible part and the easiest to estimate: monthly revenue divided by 720 hours gives the hourly figure. The search cost is slower. Repeated 5xx responses make crawlers back off and return later, but a short outage does not remove an indexed page.
What does linger is trust. A checkout that fails twice in a festival week loses customers who never file a ticket, and that loss never appears in a status page. The architectural answer, when the maths justifies it, is server clustering rather than a better promise.
How Uptime SLA Compensation Claims Work
A claim needs three things: a timestamped record of the failure, a check that the cause is not on the exclusion list, and a ticket inside the window the agreement sets. Without the first, the host can only compare its own probe data with your description.
- Export the downtime log from your monitor with start and end times in a stated timezone.
- Check the provider status page for a matching incident, and screenshot it.
- Rule out your own DNS, plugin or resource-limit suspension before filing.
- Open the ticket within the claim window and ask which credit the month qualifies for.
- Keep the written reply; the credit normally appears on the next invoice, not as a cash refund.
WebJi hosting states a 99.9% uptime target and monitors it externally; our data centres and monitoring page describes where those checks run. Ask support for the current credit terms rather than assuming a figure.
Frequently Asked Questions
What does 99.9% uptime mean?
It means the service should answer correctly for 99.9% of the measurement window, leaving 0.1% for outages. In a 30-day month that allowance is 43.2 minutes.
How much downtime is 99.9% per month?
About 43 minutes in a 30-day month of 43,200 minutes. The same target allows roughly 8 hours 46 minutes across a full year.
Is 100% uptime possible?
Not for a single server, because patching, hardware failure and network faults all interrupt service eventually. Redundancy across machines gets close, and that is what clustering and failover buy.
How do hosting companies measure uptime?
Usually with external HTTP probes that check a URL every one to five minutes from one or more regions. The percentage is therefore a sample, and a short outage between checks can be missed.
Does scheduled maintenance count as downtime?
Normally no, because SLAs exclude announced maintenance windows. Check how much notice the contract requires before the exclusion applies.
What is an uptime SLA and do I get compensation?
An SLA is the contract clause that sets the target, the measurement window, the exclusions and the remedy. The remedy is usually a service credit on a future invoice, and it must be claimed inside the stated window.
Is WebJi® an Indian company?
Yes. WebJi® is headquartered in Jaipur, Rajasthan and has operated since 2005. Billing, support and governance all run from India.
Do you provide GST invoices?
Every purchase and renewal gets a GST invoice with our GSTIN, and hosting prices are quoted exclusive of 18% GST. Claim input credit as applicable.