Service Level Agreement

Last updated: August 7, 2026

Somewhere in a Midrand office park, two IT companies are arguing about a three-line email that said “we’ll keep your systems up.” The client thinks “up” means their payroll runs at month-end no matter what. The supplier thinks “up” means the server responds 90% of the time during business hours. Neither wrote anything down, and now the dispute is costing both sides more than the original contract. That’s the gap a Service Level Agreement (SLA) exists to close. This guide explains what an SLA should contain, how to write realistic targets for the South African market, and how to avoid the common traps that turn an SLA into a source of conflict rather than clarity.

What an SLA Is and When You Need One

An SLA is the part of a service contract that defines exactly what level of service you’re buying or selling — the measurable targets, the response times, the availability guarantees, and what happens when they’re missed. You need one whenever you’re outsourcing something critical to your business: managed IT, cloud hosting, security monitoring, cleaning, transport, maintenance, call-centre services. If you’re paying someone to do something your business depends on, an SLA turns “do a good job” into “meet these specific, measurable standards.”

For a South African business, the SLA is particularly valuable because it sets expectations against the local environment. A hosting provider promising “99.99% uptime, 24/7” may be offering something they simply cannot deliver during a stage-six loadshedding event without adequate backup power. An SLA forces both sides to be honest about what’s realistically achievable — and to price in the infrastructure needed to achieve it.

The Core Components of a Good SLA

A well-written SLA is precise. Vague language is the enemy. Start with a clear description of the services covered, so there’s no argument later about whether a particular task falls within the agreement.

Measurable Service Targets

Define availability in numbers. Instead of “the system will be reliable,” write “the system will be available 99.5% of the time during business hours, measured monthly.” Define response times — “critical incidents acknowledged within 30 minutes, resolved within 4 hours.” Define resolution times for different severity levels. Every target must be specific and measurable, or it’s just an opinion.

Response and Escalation Matrix

Who do you contact, when, and what happens if the first response doesn’t fix the problem? Set out a clear escalation path with named roles and timeframes at each level. For a security company in Durban or a support desk in Johannesburg, knowing exactly who to escalate to and when prevents the classic “it’s been three days and nobody’s told me anything” scenario.

Reporting and Monitoring

How is performance measured, and who measures it? Agree on regular reports — monthly availability reports, incident logs, response-time averages — and on the measurement period. If the SLA says 99.5% availability but never defines how it’s calculated, the number is meaningless. Define it: “availability is calculated as total minutes in the month minus downtime, divided by total minutes.”

Remedies for Breach

What happens when targets are missed? SLAs typically include service credits — a partial refund or credit against future invoices when the provider misses targets. These should be meaningful enough to focus the provider’s attention but not so punitive they drive them out of business. A credit mechanism that’s fair and predictable is better than an open-ended “we’ll make it right” promise.

Writing Realistic Targets for the South African Reality

This is where South African SLAs fail most often: they borrow targets from overseas benchmarks and ignore local conditions. Power supply is the obvious one. If you’re contracting for hosting or manufacturing uptime, your SLA must address what happens during loadshedding. Does the provider have backup generators and batteries? Is the client expected to have its own UPS? What happens to the availability target when the outage is caused by Eskom rather than the provider’s failure? A good SLA explicitly carves out force majeure events and defines what’s expected of each side in a power outage, rather than leaving it to be argued later.

Security and crime are another factor. If your SLA covers physical security guarding or a delivery service, the contract should be realistic about incident response times in high-risk areas and about what the provider can and can’t guarantee. And on the client side, the SLA should be clear about your own obligations — if your premises aren’t ready, your equipment isn’t maintained, or your staff don’t give access, the provider’s obligations may be suspended. An SLA is a two-way street.

Common Pitfalls That Turn SLAs Into Disputes

The most common mistakes are all about vagueness. Unmeasurable targets (“best effort”, “reasonable time”), no defined measurement period, no escalation path, and no agreed remedy for breach. Then there’s the trap of the “zero-downtime” promise — an SLA that guarantees perfection on paper but that no provider can actually meet, which guarantees disputes rather than preventing them.

There’s also the scope creep problem. An SLA that tries to cover every conceivable service in one document becomes bloated and unenforceable. Keep the SLA focused on the services that genuinely matter to your business’s operations, and cover the rest in the main contract. And remember the SLA needs to evolve. Your business in 2026 may not have the same needs as your business in 2029. Include a review mechanism — an annual or semi-annual review where both sides update targets and prices.

Completing and Using Your SLA

Work through the template systematically. Fill in the parties and the service description first, then the targets, then the reporting and escalation, then the remedies. Be specific at every step. Where the template asks for a percentage or a time, don’t leave it blank — either insert a realistic number or consciously decide to negotiate it. The blank boxes in a template are where future disputes hide.

Who Signs and How to Enforce an SLA

An SLA is only as strong as the people behind it. On the client side, the person who signs should be the person with authority to hold the provider accountable and, ideally, the person who understands what the business actually needs from the service. On the provider side, the signatory should be someone senior enough that their signature is meaningful. Enforcing an SLA comes down to the records you keep. If you want to claim a service credit for a missed availability target, you need the monitoring data and the incident reports to prove it. In South Africa, a dispute under an SLA is a contractual dispute, so the same principles of breach and remedies apply — you’d rely on the agreed credit mechanism, or in serious cases on termination rights and claims for damages. The practical rule is simple: document everything, because an SLA you can’t prove is an SLA you can’t enforce.

The Difference Between an SLA and a Service Contract

A common point of confusion is the relationship between the main service contract and the SLA. The service contract is the broad agreement that sets out the scope, the price, the parties’ general obligations, liability and termination — the “who, what, and how much.” The SLA is typically an annexure or schedule to that contract, containing the specific, measurable service levels, the reporting requirements, and the remedies for failing to meet them. They work together. The SLA references the contract’s liability and termination terms, and the contract incorporates the SLA by reference. If you have an SLA floating around without an underlying service contract, you may have set service targets but no framework for the overall commercial relationship. For a critical service, make sure both documents exist and reference each other, so the targets have a home and the contract has teeth.

Final Thoughts

A good SLA doesn’t prevent every dispute, but it removes the disputes caused by different expectations of the same words. When both sides have signed a document that says 99.5% availability, measured monthly, with defined response times and an agreed credit mechanism, the argument shifts from “you’re just making excuses” to “let’s look at the report.” That’s a far more productive conversation. Take the time to write realistic, specific, measurable targets — and to price the infrastructure needed to meet them — and your SLA becomes a working tool for a smooth relationship, not a document that only surfaces when things go wrong.

Designing Service Metrics That Reflect Your Business

The most useful SLAs start not with a percentage but with a question: what does your business actually depend on this service for? A courier fleet operator whose drivers log in at 06:00 every morning cares far more about availability during the 06:00–08:00 window than about the service being up at 02:00. An online retailer cares about response time at month-end when sales peak. Draft your targets around those peaks. A single flat “99.5% availability” figure across the whole month can hide a catastrophic outage at exactly the wrong hour.

Put real numbers on it. If you commit to 99.5% uptime, measured over a calendar month of 720 hours, that permits a maximum of roughly 3.6 hours of downtime in the month. Over a year that is about 43.8 hours. Before you sign, ask whether the provider’s infrastructure and support staffing can honestly deliver that — and, just as importantly, whether your own premises, connectivity, and staff behaviour (delaying maintenance access, failing to notify the provider of changes) will let them deliver it. An SLA is not a one-way demand; it is a contract about the conditions under which a promised level can actually be achieved.

Service Credits, Penalties and Review Cycles

When a target is missed, the usual remedy is a service credit: a reduction on the next invoice calculated on a formula, rather than an open-ended claim for damages. Agree the formula in the SLA so a breach resolves without a dispute. For example, if monthly availability falls below the agreed 99.5%, the credit might be 5% of the monthly fee for each full 1% shortfall. If the monthly fee is R20,000 and availability drops to 97%, that is a 2.5% shortfall, producing a credit of roughly R2,500 on the next invoice. Cap the credits (often at a percentage of the fee) so the remedy stays proportionate and the provider remains in business to keep serving you.

Finally, build in a formal review cycle — at least annually, and ideally quarterly for the first year. Business needs change: you may add branches, migrate workloads, or shift to a new product, and a target that was right in year one may be over- or under-specified by year two. A review clause that lets both sides revisit targets, measurement methods, and pricing keeps the SLA aligned with reality instead of becoming a stale document that no one believes. Schedule the review in the SLA itself, so it actually happens rather than being forgotten until something breaks.

Frequently Asked Questions

What is Service Level Agreement?

Service Level Agreement 👁️ Preview & Download

How does Service Level Agreement work?

The guide above walks through it step by step, with practical examples and South African context so you can apply it correctly.

Why is Service Level Agreement relevant in South Africa?

Because the details matter locally — from local rules and rates to everyday usage — this guide is written specifically for South African readers.