MKB Juristen drafts custom legal documents
It is best not to cobble together or copy important contracts, terms and conditions, and other legal documents yourself. We help entrepreneurs on a budget with customized legal solutions, clear costs upfront, and practical explanations.
- Custom contracts, terms and conditions, and legal documents
- Budget-friendly and clear about the costs upfront
- Request a free consultation or a no-obligation quote
A Service Level Agreement (SLA) defines the service level an IT supplier guarantees and what happens if that standard is not met. Key elements include: availability (uptime, for example 99.5% or 99.9%), response and resolution times per priority, service credits or penalties for underperformance, maintenance windows, and the scope of support. An SLA is often separate from the main contract, allowing service levels to be adjusted without reopening the entire contract. Below is what an SLA covers, which metrics matter, and why you might want to keep it separate.
The short answer
- Uptime: guaranteed availability, for example 99.5% or 99.9%.
- Response time: how quickly the supplier responds per priority.
- Resolution time: target time to resolve the problem.
- Service credits: discount or penalty if the standard is not met.
- Maintenance windows: planned downtime that does not count.
- Scope: what support does and does not cover.
What does an SLA regulate?
An SLA translates “the service must work well” into measurable agreements. Without figures, a service promise is unenforceable: “we strive for high availability” has little legal significance. With an SLA, you know exactly what availability is guaranteed, within what timeframe disruptions will be responded to, and what you receive in return if that fails.
Uptime and availability
Availability is expressed as a percentage of the time the service is available. The figures seem close, but the difference is significant:
- 99%: over 3.5 days of downtime per year.
- 99,5%: approximately 1.8 days per year.
- 99,9% (“three nines”): approximately 8.8 hours per year.
- 99,99%: approximately 52 minutes per year.
Pay attention to the details: is availability measured per month or per year, and is planned maintenance included? A percentage without a measurement period and without a definition of “downtime” is extremely weak.
Response time and resolution time per priority
A good SLA splits incidents into priorities and assigns separate timeframes to them. Response time is how quickly the supplier responds; resolution time is the target time to resolve the problem. An example breakdown:
- P1 – critical (service down): response within 1 hour, solution as soon as possible.
- P2 – high (important function disabled): response within 4 hours.
- P3 – normal (disruption, workaround possible): response within 1 working day.
- P4 – low (question or request): response within a few working days.
Important distinction: response time is usually strictly enforceable, whereas resolution time is often a best-efforts obligation. Also establish support hours — 24/7 or office hours only — because that determines whether that “1 hour” applies at night as well.
Service credits and fines
A standard without consequence is no standard. Service credits are the most common sanction: if the supplier fails to meet the agreed uptime, you receive a discount on the invoice, often increasing with the severity of the delay. For example, a 5% credit for availability below 99.5%, or 10% below 99%.
What to look out for:
- Automatic or on demand: do you have to claim credits yourself, or are they automatically deducted?
- Ceiling: suppliers often cap credits at a percentage of the monthly fee.
- Exclusivity: does it state that credits are the “only remedy”? If so, you may be excluding other damages.
For business-critical services, you want credits to be an incentive, not a buyout fee that excludes further liability.
Maintenance windows and scope
Scheduled maintenance normally does not count as downtime, provided it falls within an agreed window (for example, Sunday night) and is announced in advance. Specify how far in advance and how often. This prevents a supplier from “meeting” their uptime target by labeling everything that goes wrong as maintenance.
The support scope determines what is and is not covered: which components, which channels (telephone, ticket, email), and what explicitly falls outside of support (customization, third-party integrations, user errors).
Why the SLA separate from the main contract
Service levels change faster than the rest of the agreements. By keeping the SLA as a separate appendix, you can adjust uptime, priorities, or support times without renegotiating the entire contract. However, do include a precedence clause in the main contract so that it is clear which document takes precedence if the SLA and the general terms and conditions conflict.
Practical example
An administration office relies entirely on an online accounting package. The SLA stipulates 99.9% monthly availability, a response within one hour in the event of a critical outage during business hours, and a 10% service credit if uptime drops below 99.5%. When the package was unavailable for three hours on a peak day during the tax return period, this fell well below the standard, and the credit was automatically deducted from the next invoice. Without the SLA, the office would not have been able to claim anything.
Honest recommendation
For a non-critical standard tool, the vendor's SLA is usually sufficient — pay particular attention to how uptime is measured and whether credits are automatically generated. You don't need a lawyer for that. As soon as the service is business-critical, have the SLA reviewed: pay attention to the measurement period, the definition of downtime, whether credits are the only legal remedy, and the credit ceiling. A strict SLA on a service your business depends on is well worth the cost.
Read more about our SLA setup, the overview of IT contracts , and the SaaS agreement that often includes an SLA.
Frequently Asked Questions
A service level agreement specifies the service level an IT supplier guarantees: availability (uptime), response and resolution times per priority, maintenance windows, and the consequences (service credits) if the standard is not met.
99.9% availability represents approximately 8.8 hours of downtime per year. For comparison: 99.5% is approximately 1.8 days and 99.99% is approximately 52 minutes per year. Note whether it is measured per month or per year and whether scheduled maintenance is included.
Service credits are a discount or penalty you receive if the supplier fails to meet the agreed service levels, for example, 5% of the monthly fee for uptime below 99.5%. Note whether they are automatically deducted and whether they are the only legal remedy.
Response time is how quickly the supplier responds to a report; resolution time is the target time to resolve the problem. Response time is usually strictly enforceable, whereas resolution time is often a best-efforts obligation. Both are recorded per priority.
Usually not, provided it falls within an agreed maintenance window and is announced in advance. Specify how far in advance and how often, so that a supplier cannot meet their uptime target by labeling outages as maintenance.
Service levels change more frequently than the rest of the agreements. As a separate appendix, you can adjust uptime, priorities, or support times without reopening the entire contract. A ranking clause determines which document takes precedence in the event of a dispute.
For a non-critical standard tool, usually not. As soon as the service is business-critical, however: have the measurement period, the definition of downtime, the credit ceiling, and the question of whether credits are the only legal remedy reviewed.
