Uptime SLAs: What Your Hosting Provider Actually Promises

What Is an Uptime SLA?

An SLA — Service Level Agreement — is a contract between you and your hosting provider. It specifies the minimum percentage of time your website or service should be available. If the provider fails to meet that target, you're entitled to service credits.

That's the theory. In practice, most SLAs are written to protect the provider, not you.

Hosting companies advertise "99.9% uptime guarantee" like it's a promise etched in stone. But when you actually read the SLA document, you'll find it's more like a suggestion with a long list of exceptions. Understanding what your SLA actually covers — and what it doesn't — is the difference between having a real guarantee and having a marketing slogan.

How to Read an Uptime SLA

Every hosting SLA has three parts that matter: the uptime commitment, the measurement method, and the credit schedule. Most people only look at the first one. That's a mistake.

The uptime commitment

This is the headline number — 99.9%, 99.95%, 99.99%. It tells you the maximum amount of downtime the provider considers acceptable in a given period (usually monthly).

The measurement method

This is where things get interesting. How does the provider measure uptime? Some count only their core infrastructure. Others measure from the network edge. A few measure from the customer's perspective. The measurement method determines what counts as "down."

Watch for language like "server availability" versus "service availability." Server availability means the physical machine is powered on. Your website could still be unreachable due to a networking issue, a misconfigured load balancer, or a DNS problem — and none of that counts as downtime under a server-availability SLA.

The credit schedule

This defines what you get when the provider misses their target. It's almost never a cash refund. It's a credit toward future service — meaning you have to keep paying them to use it.

Typical credit schedules look like this:

Notice that even if your site is down for an entire day (96.7% uptime for the month), you're getting a 25% credit — not a full refund, and certainly not compensation for lost revenue. The credit never exceeds 100% of your monthly hosting bill, regardless of how much money you lost.

Example SLA Language, Dissected

Here's a paragraph you might find in a typical hosting provider's SLA. Let's break down what it actually means:

"Provider guarantees 99.9% monthly uptime for Dedicated Server hardware and network availability, excluding scheduled maintenance windows, customer-initiated actions, force majeure events, and third-party service disruptions. Downtime is measured from the time a trouble ticket is opened by the customer until service is restored."

There are four things buried in this paragraph that dramatically limit your coverage:

1. "Dedicated Server hardware and network availability" — This only covers the physical server and the provider's own network. If the problem is with their software stack, their control panel, or any shared service, it doesn't count.

2. "Excluding scheduled maintenance windows" — The provider can take your site offline for maintenance and it won't affect their uptime number. Some providers give 24 hours' notice. Others give an hour. Some define "scheduled" very loosely.

3. "Force majeure events" — Natural disasters, sure. But some SLAs stretch this to include power grid failures, government actions, or even "circumstances beyond our reasonable control" — which can mean almost anything.

4. "Downtime is measured from the time a trouble ticket is opened" — This is the big one. Your site could be down for two hours before you notice. If you don't open a ticket during that time, those two hours don't count. The clock only starts when you report it.

Common SLA Exclusions

Every hosting SLA has exclusions. Here are the ones that catch most small business owners off guard.

Scheduled maintenance

Providers reserve the right to perform maintenance — and that maintenance window is excluded from uptime calculations. Some providers schedule maintenance weekly. Others do it monthly. A few do it whenever they want with minimal notice.

The key question: How much notice do they give, and how long can maintenance last? Some SLAs don't cap maintenance windows at all.

DDoS attacks

Most SLAs exclude downtime caused by distributed denial-of-service attacks. From the provider's perspective, the infrastructure is working fine — it's just overwhelmed by malicious traffic. This is arguably the most common cause of unexpected downtime for small businesses, and it's almost never covered.

"Acts of God" and force majeure

Natural disasters, power outages, fires, floods. Reasonable enough. But read the full list carefully. Some providers include "labor disputes," "government actions," "Internet disturbances," or the catch-all "events beyond our control." That last one could cover just about anything.

Network issues outside the provider's control

If the problem is between your visitors and the provider's data center — an ISP outage, a submarine cable cut, a BGP routing issue — that's not the provider's fault. Fair enough, but it means your site can be unreachable to a significant portion of your audience and the SLA is technically unbroken.

Customer-caused issues

If your code crashes the server, if you misconfigure something, if your CMS gets hacked — that's on you. No SLA credit. This is reasonable, but the line between "customer-caused" and "provider-caused" can be blurry. A hosting platform that auto-updates and breaks your site might argue that's a customer configuration issue.

How to Claim SLA Credits

Here's the uncomfortable truth: most businesses never claim the credits they're owed. Providers know this. It's why they can offer generous-sounding SLAs — the redemption rate is extremely low.

Step 1: Document the downtime

You need evidence. Screenshots, logs, timestamps. If your SLA requires you to open a support ticket for downtime to "count," you need to open that ticket immediately — even at 3 AM.

This is nearly impossible to do consistently without automated monitoring. If you're asleep when your site goes down at midnight and it comes back up at 2 AM, you have no ticket and no claim.

Step 2: File within the claim window

Most SLAs require you to submit credit requests within a specific timeframe — often 30 days of the incident. Miss the window and you forfeit the credit, no matter how severe the outage was.

Step 3: Provide the required documentation

Providers typically require the exact times of the outage, the affected services, and sometimes proof that the downtime was not caused by any excluded category. Without independent monitoring data, you're relying on the provider's own logs — which conveniently might not show the problem.

Step 4: Wait

Credit requests are reviewed by the provider. They decide whether the downtime qualifies. They can reject your claim if they determine it falls under an exclusion. You may need to escalate or dispute their decision.

Why You Need Independent Monitoring

Your hosting provider is both the service provider and the one measuring whether they've met their SLA. That's a conflict of interest.

Independent monitoring solves this. A third-party tool checks your site from multiple locations at regular intervals and records every outage — including the ones that happen at 3 AM, the ones that only affect certain regions, and the ones your provider's dashboard conveniently doesn't report.

With independent monitoring, you get:

• Timestamped proof of downtime that the provider can't dispute
• Alerts the moment your site goes down so you can open a support ticket immediately (starting the SLA clock)
• Monthly uptime reports you can compare against the provider's own numbers
• Historical data showing patterns — maybe your provider has maintenance windows every Tuesday night that add up to hours of downtime per year

Without independent monitoring, you're trusting the provider to honestly report their own failures. Some do. Many don't.

What a Good SLA Looks Like

Not all SLAs are designed to avoid accountability. Here's what to look for in a provider that actually stands behind their uptime promise:

• Measurement from the customer's perspective, not just server availability
• Automatic credits applied without requiring you to file a claim
• Short or no exclusion list — ideally no maintenance-window exclusion, or a guaranteed maximum maintenance window per month
• Cash refund option, not just service credits
• Independent third-party monitoring referenced in the SLA itself

These are rare. But they exist, particularly among premium and enterprise-grade hosting providers. If your hosting provider offers none of the above, their SLA is more of a marketing tool than a guarantee.

The Bottom Line

An uptime SLA is better than no SLA. But it's not a safety net — it's a minimum standard with significant loopholes. The credit you receive when your provider misses their target rarely comes close to covering the actual cost of downtime to your business.

The real protection isn't in the SLA document. It's in having independent monitoring that catches outages in real time, documented proof of every minute of downtime, and the ability to hold your provider accountable — or switch to one that actually delivers.

Start by understanding what your current SLA actually promises. Use the SLA Uptime Calculator to see what your provider's uptime percentage means in real downtime minutes. Then set up independent monitoring so you know — not guess — whether they're meeting their commitment.

---

Part of Boring Tools — boring tools for boring jobs.