Why Your ISP's SLA Doesn't Mean What You Think It Does

Somewhere in your office there is a contract with your internet service provider. It contains a section called an SLA — a Service Level Agreement. It probably promises 99.9% uptime, a response time if something goes wrong, and a credit if they miss the target.

It almost certainly does not protect you the way you think it does.

This is not unique to Ghana. It is how the ISP industry works globally. But in a market where network reliability is already a challenge, understanding exactly what your SLA does and does not cover is not a legal technicality — it is a business decision.

What a Typical ISP SLA Actually Covers

Most ISP SLAs cover one thing: whether the physical connection between their network and your premises is up or down. This is called the demarcation point — the point where their responsibility ends and yours begins.

If the fibre coming into your building is active, the SLA is satisfied. What happens after that — inside your building, across your local network, through your switches and access points, all the way to the application your staff are trying to use — is outside the SLA entirely.

This matters because the majority of connectivity problems businesses experience are not last-mile failures. They are performance degradation, not outages. Slow speeds during peak hours. Video calls that stutter. VoIP that drops. Cloud applications that time out. None of these necessarily trigger the SLA because the link itself is technically "up."

The 99.9% Uptime Promise

99.9% uptime sounds excellent. In practice, it means your provider is contractually permitted to have your connection down for approximately 8.7 hours per year without breaching the agreement.

That is more than one full business day of downtime — spread across the year in ways that are difficult to predict and expensive to absorb.

More importantly, "uptime" in most ISP contracts is measured at the network level, not at the experience level. A connection that is technically up but running at 10% of its contracted speed for six hours is not a breach of most SLAs. You experienced a terrible day of connectivity. Your provider met their contractual obligation.

What the Fine Print Usually Contains

Every ISP SLA contains exclusions. The specific language varies, but common exclusions include:

Force majeure. Events outside the provider's control — power outages, extreme weather, civil unrest — are typically excluded. In Ghana, where power instability is a real operational consideration, this exclusion can be significant.

Scheduled maintenance. Most SLAs exclude downtime caused by planned maintenance windows, often defined broadly. A provider can schedule maintenance at 2am and the downtime does not count against the SLA.

Customer-side equipment. If the fault is determined to be on your side of the demarcation point — your router, your switches, your cabling — the SLA does not apply. The definition of "customer-side" can be broader than you expect.

The credit structure. Even when an SLA breach is confirmed, the remedy is typically a credit against your next bill — not compensation for lost business. A credit of one day's service fee does not come close to covering the cost of a day of downtime for most businesses.

Why Monitoring Direction Matters

Here is the core issue: how does your provider know there is a problem?

Most ISPs monitor from their Network Operations Centre (NOC) to your demarcation point. That tells them the last mile is up. It tells them nothing about what you are actually experiencing inside your building, or how your connection performs across the broader internet — across backhaul, peering, and the real servers your business depends on.

If your connection is degraded — not down, but slow, high-latency, dropping packets — the NOC may see nothing wrong. You will know something is wrong. Your staff will know. Your customers may know. But the system that triggers SLA accountability is not measuring what you are experiencing.

This is why the monitoring direction matters. Monitoring from the customer premises to a real internet endpoint — not from the NOC to the demarcation — captures what the customer actually experiences. Latency, packet loss, jitter, and real throughput. Not just whether the link is up.

What a Real SLA Looks Like

A contractually enforceable SLA requires three things: comprehensive monitoring, a clear definition of what is being measured, and a provider who is accountable for the full experience — not just the last mile.

Comprehensive monitoring means measuring from inside your premises outward, continuously, to a real internet endpoint. Not a local speed test server. Not a CDN node. The actual internet.

A clear definition means specifying not just uptime but quality metrics: maximum acceptable latency, packet loss thresholds, and minimum sustained throughput. These are measurable. They can be contractually binding.

Accountability means the provider reacts when the monitoring alerts — not when you call to complain. The difference between a proactive managed service and a reactive ISP is the difference between a problem resolved before you notice it and a problem that has already cost you half a day of productivity.

Questions to Ask Your Current Provider

Before you sign a renewal — or when you are evaluating a new provider — ask these questions directly:

• Where exactly does your monitoring occur? At the NOC, at the demarcation, or inside my premises?
• What specific metrics does the SLA cover? Uptime only, or latency, packet loss, and throughput?
• How is downtime defined? Is a degraded connection that remains technically "up" covered?
• What are the exclusions? Ask for the full list in plain language.
• What is the remedy if you breach? Credit only, or genuine compensation?
• Can I access your monitoring data in real time?

A provider who cannot answer these questions clearly, or who becomes evasive when you ask them, is telling you something important.

The IJA Approach

IJA Technologies monitors from a device installed at your premises — a Raspberry Pi running continuous quality tests — all the way to an AWS server in London. Every test crosses the last mile, the backhaul, and the peering fabric to reach the real internet.

We measure latency, packet loss, jitter, and true internet throughput on a continuous schedule. Every customer receives access to their own Grafana dashboard showing exactly what we see, in real time.

Our SLAs are contractual because our monitoring is comprehensive. We react when our systems alert us — not when you call to complain.

That is what a real SLA looks like.

IJA Technologies is Ghana's carrier-agnostic managed technology partner. We design, build, and manage the technology infrastructure that keeps businesses connected, secure, and visible. Talk to us about your network.