The Uptime / SLA Clause Explained (2026)

An uptime clause, usually part of a service level agreement (SLA), is the promise in a software or hosting contract that the service will be available for a stated percentage of time, with a defined remedy if the vendor falls short. It is the clause that turns "the service should work" into a measurable commitment with consequences. For any business that depends on a SaaS or cloud service, it is one of the most important clauses in the contract. This page explains what the clause does, how the numbers translate into real downtime, how credits and exclusions work, and what each side should negotiate. It is general information, not legal advice. ## What does an uptime / SLA clause do? It does four things: 1. Sets a target. An availability figure, usually expressed per calendar month, such as 99.9 percent. 2. Defines how it is measured. What counts as "available", what counts as "downtime", and the exact formula. 3. Lists exclusions. Periods that do not count against the vendor, such as scheduled maintenance. 4. Sets a remedy. What the customer gets if the target is missed, usually service credits, and whether that is the only remedy. All four matter. A high headline percentage means little if the definition of "available" is narrow, the exclusions are broad, or the only remedy is a small credit the customer has to chase. ## What the percentages actually mean The headline number translates into a surprising amount of allowed downtime. Rough monthly and yearly figures: | Uptime target | Allowed downtime per month | Allowed downtime per year | |---|---|---| | 99 percent ("two nines") | about 7.3 hours | about 3.65 days | | 99.9 percent ("three nines") | about 43.8 minutes | about 8.8 hours | | 99.95 percent | about 21.9 minutes | about 4.4 hours | | 99.99 percent ("four nines") | about 4.4 minutes | about 52.6 minutes | | 99.999 percent ("five nines") | about 26 seconds | about 5.3 minutes | So a 99.9 percent SLA, which sounds close to perfect, still permits roughly three quarters of an hour of downtime every month. Decide whether that is acceptable for how critical the service is to you. ## How uptime is calculated The formula usually looks like this: > Uptime percentage = (total minutes in the month minus downtime minutes) divided by total minutes in the month, expressed as a percentage. The definitions around the formula decide how generous it really is: - What is "downtime"? Total unavailability, or only errors above a threshold? Does a slow but working service count? Partial outages?

• How is it detected? Vendor monitoring only, or can the customer's own evidence count?
• Over what period? Per calendar month is most common; a longer averaging window hides short, severe outages. ## Exclusions: where the promise shrinks Most SLAs carve out periods that do not count against the vendor. Common exclusions: - Scheduled maintenance, ideally within a notified window and capped in hours.
• Emergency maintenance.
• Force majeure and events beyond the vendor's control.
• Third-party failures, such as the customer's own network or a utility outage.
• Beta or trial features.
• Customer-caused issues, such as misuse or exceeding limits. Watch for exclusions so broad they swallow the commitment, for example uncapped scheduled maintenance, or a blanket "any third-party service" when the vendor itself relies on major cloud providers. ## Service credits and the sole-remedy trap The usual remedy is a service credit: a percentage of the monthly fee, refunded or applied to a future invoice, scaled to how badly the target was missed. Typical features: - Credits tiered by severity, often in the range of 5 to 30 percent of the monthly fee, rising as uptime falls.
• An overall cap, commonly up to 100 percent of one month's fee.
• A claim requirement: the customer must request the credit within a short window (sometimes days), or lose it.
• Often a sole and exclusive remedy clause: credits are the only thing the customer can claim. The sole-remedy wording is the trap. If credits are the only remedy, a customer may be unable to terminate or claim damages even for repeated, serious outages. Negotiate a termination right for chronic failure (for example, missing the target in several consecutive months, or a single prolonged outage) so the credit is not the end of the story. ## A sample uptime clause > The Provider will use commercially reasonable efforts to make the Service available with a Monthly Uptime Percentage of at least 99.9 percent during each calendar month. "Monthly Uptime Percentage" means total minutes in the month minus Downtime minutes, divided by total minutes in the month. "Downtime" means any period in which the Service is unavailable, excluding Scheduled Maintenance (notified at least 48 hours in advance and not exceeding 4 hours per month), emergency maintenance, force majeure, and failures of the Customer's own equipment or network. If the Provider fails to meet the target in a given month, the Customer may claim a service credit, requested in writing within 30 days, calculated as: 10 percent of the monthly fee for uptime below 99.9 percent, 25 percent below 99.0 percent, and 50 percent below 95.0 percent, up to a maximum of 100 percent of that month's fee. If the Provider fails to meet the target in three consecutive months, the Customer may terminate the affected Service on written notice. This is an illustration, not a model to copy without advice. Figures and definitions should match the service and the risk. ## What each side should negotiate As a customer, push for: a target that fits how critical the service is; a clear, fair definition of downtime; capped and notified maintenance windows; a sensible claim window; and a termination right for chronic failure on top of credits. As a vendor, protect: realistic targets you can actually meet; clear exclusions for things outside your control; a credit cap; and a claim process so credits are predictable. Promising five nines you cannot hit is worse than promising three nines you can. ## Common mistakes - Reading the headline percentage and ignoring the definition of "available" and the exclusions.
• Accepting credits as the sole remedy with no termination right for repeated outages.
• Missing the claim window, so earned credits are forfeited.
• Allowing uncapped scheduled maintenance that quietly erases the commitment.
• Not aligning the uptime clause with the liability and termination clauses elsewhere in the contract. ## Seeing every SLA commitment in one place An uptime clause is easy to read once and forget. The hard part is remembering, across every vendor contract, which SLA you agreed to, what the credit tiers are, how narrow the exclusions were and when a claim window opens after an outage. The work of drafting SLA terms from a house standard, flagging a weak exclusion or a sole-remedy trap, keeping versions straight and surfacing the dates and thresholds that matter is the work an AI contract workspace can hold in one place: draft from your own template, flag clauses that drift from your standard, keep a live timeline of changes, and raise key obligations as alerts. If that is useful, Adira is built around that single-workspace idea. This article is general information about SLA drafting practice and is not legal advice. For a specific contract, take qualified advice in the relevant jurisdiction.