Uptime is the percentage of time your cloud server stays active and reachable. A responsible SLA (Service Level Agreement) should guarantee at least 99.9% monthly availability — fewer than 44 minutes of downtime per month.
If your website, online store, or application runs on cloud hosting, every minute of downtime has a real cost: lost sales, damaged reputation, and frustrated users. Signing up with a provider without reviewing their SLA is one of the most common — and most expensive — mistakes small businesses make.
This article explains what uptime percentages actually mean, how to read SLA clauses, and what minimum guarantees you should demand before signing any cloud hosting contract.
What Does Uptime Mean and How Is It Calculated?
Uptime is expressed as a percentage of total time in a given period (usually a 30-day month = 43,200 minutes). Here's how the numbers translate into real downtime:
| Guaranteed Uptime | Downtime/Month | Downtime/Year |
|---|---|---|
| 99.0% | ~7 h 12 min | ~3 days 15 h |
| 99.5% | ~3 h 36 min | ~1 day 20 h |
| 99.9% | ~43 min | ~8 h 46 min |
| 99.95% | ~21 min | ~4 h 23 min |
| 99.99% | ~4 min | ~52 min |
The gap between 99.0% and 99.9% looks tiny on paper, but it's nearly 7 extra hours of downtime per month. For an active e-commerce store, that can mean thousands of dollars in lost revenue.
What a Cloud Hosting SLA Must Include
An SLA is more than a single percentage. Before signing, verify that the contract explicitly addresses each of the following:
Definition of "downtime"
Some providers only count outages where the server is completely unreachable. Others exclude incidents shorter than 5 minutes, scheduled maintenance windows, or failures caused by third parties (DNS, CDN). Demand that the SLA spell out exactly what counts as downtime.
Measurement methodology
The provider must state where and how often they monitor availability. Top-tier services use external probes from multiple geographic locations every 60 seconds or less — not internal tools that only catch total node failures.
Credits and compensation
When the provider misses the SLA target, what happens? Look for automatic service credits applied to your next invoice, proportional to the actual downtime. Avoid contracts where you must "open a ticket and wait for approval" to receive a credit — that process typically costs more in your time than the credit is worth.
Exclusions and force majeure
Read the exclusions list carefully. Excluding massive DDoS attacks or regional Internet failures is reasonable; routinely excluding the provider's own scheduled maintenance is not.
How Much Uptime Does Your Business Actually Need?
The answer depends on the economic impact of each minute offline. Consider these scenarios:
- Blog or informational site: 99.5% may be acceptable. An hour of downtime per month rarely affects direct revenue.
- Corporate website with contact forms: 99.9% is the reasonable minimum. A drop during a Google Ads campaign can cost more than a full year of hosting.
- Active online store: 99.95% or higher. Every minute of downtime during peak season (Black Friday, flash sales) has a direct cost in lost orders.
- SaaS app or platform with concurrent users: 99.99% ("four nines"). Here the SLA is just the starting point — you also need your own redundant architecture.
If your business operates on specific schedules (say, Monday–Friday 9 a.m.–6 p.m.), the real impact of an overnight outage is lower. That said, search engine crawlers work around the clock, and a prolonged outage can hurt your SEO rankings.
Red Flags in an Unreliable SLA
Not every SLA is written in good faith. Watch for these warning signs before you commit:
- The SLA promises 99.9% uptime, but the exclusions are so broad that the remaining 0.1% always falls within them.
- Credits are capped at a low amount (e.g., a maximum of 5% of the monthly fee regardless of how long the outage lasted).
- No public incident history or real-time status page exists.
- The contract requires you to prove the outage with your own screenshots and logs.
- 24/7 support is not included — an incident at 2 a.m. Saturday can stretch until Monday if the technical team is unavailable.
Before signing, search for the provider's name alongside "status page" or "incidents" to review their outage history. Reliable providers publish it openly; those who don't are hiding something for understandable reasons.
How to Improve Availability Beyond the SLA
A solid SLA is necessary but not sufficient. Your platform's real-world availability also depends on decisions you control:
- Active CDN: A content delivery network serves your site from global caches even during short origin outages.
- External monitoring: Tools like UptimeRobot (free) or Better Uptime alert you within seconds and generate independent logs you control.
- Off-site backups: If your provider suffers a major incident, you need to restore somewhere else. Backups stored in the same data center won't help if that data center is down.
- Application-level health checks: A server that's "up" but returning 500 errors passes a ping monitor but fails your users. Build a
/healthendpoint that validates your database connection, cache layer, and critical services.
If you need expert guidance selecting cloud infrastructure and negotiating solid SLAs, elenlace.com helps SMBs and mid-size companies build reliable environments from day one.
Explore more articles on this topic in our cloud hosting category.
Key Takeaways
- Uptime is measured as a monthly percentage; the gap between 99.0% and 99.9% equals nearly 7 extra hours of downtime per month.
- A real SLA must define what counts as downtime, how it's measured, and what compensation applies automatically.
- For online stores or mission-critical apps, the acceptable minimum is 99.95%; for informational sites, 99.9% usually suffices.
- Broad exclusions and low credit caps turn an apparently strong SLA into an empty promise.
- Complement your provider's SLA with your own external monitoring, a CDN, and backups outside the same infrastructure.
Ready to move to a cloud hosting provider with a transparent SLA? The team at elenlace.com can help you evaluate proposals, compare contracts, and launch your environment with the availability standards your business requires.
FAQ
What happens if my provider misses the SLA?
You are entitled to the credits or compensation stated in the contract. Open a support ticket citing the incident with exact dates and timestamps. If the provider refuses without justification, that is sufficient reason to switch providers.
Does 99.9% uptime include scheduled maintenance?
It depends on the contract. Many providers exclude planned maintenance windows from the uptime calculation. Read the exclusions section before signing, and ask how frequently and for how long maintenance windows typically occur.
How can I verify whether my site was down?
Use an external monitoring tool such as UptimeRobot, Freshping, or Better Uptime. These services check your URL every minute from external servers and log every incident with start and end times — independently of what your provider reports.
Does the SLA cover only the web server, or also email and the database?
Typically the SLA covers web server accessibility (HTTP/HTTPS). Email and database services may have separate SLAs or be excluded entirely. If your business depends on both, request specific SLAs for each service before signing.
Further reading
Other providers and guides worth comparing:
