Learn what an uptime guarantee means legally and why 99.9% availability allows about 43 minutes and 50 seconds of downtime per month...
An uptime guarantee is a contractually assured minimum availability of an IT service, expressed as a percentage. At 99.9%, the calculated downtime budget is about 43 minutes and 50 seconds per month, based on an average month. Whether this value is truly binding depends solely on the wording: a guarantee creates an obligation, a mere target does not.
In short:
- For a legally reliable commitment, clear measurement methods, defined reference periods, and concrete consequences for falling short are essential.
- Several short outages are perceived as more disruptive by users than a single longer outage, which underlines the importance of the time frame.
- If the value is not met, you should rely on precise evidence such as monitoring logs and ticket histories to enforce claims.
- Technically, redundant infrastructure, failover systems, monitoring, and DDoS protection secure the actual availability of a hosting platform.
A service level agreement (SLA) is not a standalone type of contract with fixed statutory rules, but part of the hosting or IT contract. It specifies what the provider actually owes, and that is exactly where its significance lies. Under the principle of freedom of contract, both sides can negotiate availability values freely, but for pre-formulated clauses, the review of general terms and conditions under Sections 305 et seq. of the German Civil Code (BGB) applies. Unclear or surprising clauses are interpreted against the party that introduced them, which for hosters is usually the provider itself, as a specialist article on SLA agreements explains.
The distinction between a guarantee and a target is decisive. Wording such as “the provider warrants an availability of 99.9%” or “guarantees” creates a genuine obligation. Sentences such as “the provider aims for an availability of 99.9%” or “endeavors to” are, by contrast, non-binding statements of intent with no enforceable effect. Anyone reviewing a contract should look closely at this wording, because the difference between the two variants is the difference between a right to remedy and a mere promise without consequences.
Several elements increase the legal enforceability of a commitment:
Without these components, even an impressive-sounding percentage often has no consequences in practice, because the customer lacks the evidence in a dispute.
An SLA clause that truly holds up in a dispute consists of more than a single percentage. A guide to contract design for cloud computing recommends explicitly regulating the following points in the contract or the general terms and conditions:
An article on SLA agreements stresses that unclear wording on these points is what most often leads to disputes in practice, because both sides have different expectations of the same percentage.
Pro tip: Ask for a written definition of the measurement method before signing the contract. A provider that cannot answer this question clearly probably has no reliable internal monitoring either.
The formula behind it is simple: downtime equals (1 minus the uptime percentage) multiplied by the reference period. Anyone who only reads the percentage often underestimates how differently individual tenths and hundredths of a percent affect the result.
The numbers make it clear: The jump from 99% to 99.9% reduces monthly downtime from over seven hours to under one hour, which is the real lever when choosing a provider.
The reference period itself is also important. An SLA uptime calculator points out that monthly SLAs appear stricter with recurring short outages, while annual values can more easily “hide” a single long outage. For an online shop with daily revenue, an outage on the first of the month is more business-critical than for a game server that is used in the evening. For live game servers, it is not just raw reachability that counts but also whether the service stays stable during peak hours, because a technically “reachable” server with high latency still feels like an outage to players.
If the guaranteed availability is not met, the specific legal consequence depends heavily on what the contract provides:
An article on SLA agreements makes clear that service credits are often worth little economically if the liability limits in the contract are set low anyway. A credit of one daily rate is of little use if the outage caused many times that amount in damage.
Evidence is decisive for enforcement. Anyone who wants to assert claims needs monitoring logs, complete ticket histories, and ideally formal incident documentation modeled on an RFC process. Without this evidence, even a justified claim remains hard to enforce in practice, as an older specialist article on service level agreements already notes. In practice, a staged approach is advisable: first report the disruption to the provider in writing and set deadlines, then escalate if there is no response, and only involve a lawyer in the case of significant financial damage or repeated violations.
A percentage in the contract is only as credible as the technology behind it. In its material on high availability, the BSI (German Federal Office for Information Security) describes availability, reliability, and maintainability as interrelated criteria that cannot be achieved through redundant hardware alone but also require mature organizational processes such as incident management and change management.
On the technical side, several building blocks contribute to actual availability:
For game servers and other latency-sensitive applications, another layer comes into play. An SLA that only measures raw reachability captures neither packet loss nor latency spikes, which Nexthosting treats as a quality dimension of its own alongside availability. A server that is “online” but runs with high latency or stuttering still feels like an outage to players, even if the SLA is formally met.
DDoS protection is also part of the basics, because a volumetric attack can take down an otherwise redundantly designed service within seconds. In addition, regular, well-communicated maintenance windows and automated backups reduce the risk of unannounced outages. Performance optimization also plays a role: for data-intensive services such as online shops, for example, using caching plugins can reduce server load and thus indirectly contribute to stability.
Pro tip: Ask a provider specifically about the difference between internal and external monitoring. A provider that uses both and makes the results accessible to the customer signals genuine confidence in its own infrastructure.
The abstract percentage only becomes tangible in a concrete case.
Many disputes arise not from malicious providers but from unclear expectations on both sides.
In my view, clarity pays off for both sides. A precise SLA protects the customer from empty promises, but it also gives the provider a realistic framework against which it can be measured, instead of hiding behind vague wording. Providers that transparently document their own downtime and disclose monitoring data on request deserve more trust in practice than those that simply put a high percentage on their marketing page. Anyone operating a game server or shop who is evaluating a commitment should treat exactly this willingness to be transparent as the signal, not the number itself.
— Erik
If you apply the criteria above to a hosting contract, you will find at Nexthosting an infrastructure with NVMe SSD storage, integrated DDoS protection at no extra charge, and a data center location that enables GDPR-compliant data processing. For multiplayer projects such as Minecraft, FiveM, or Garry’s Mod, stable performance under load matters most alongside pure reachability, and the infrastructure is designed for exactly that. For technical questions about server configuration or availability, direct support is available through a ticket system.
If you are looking for a game server with predictable performance, you will find suitable plans in the game server category, for example Minecraft from €5.57 per month or FiveM from €6.14 per month. For projects with higher demands on resources and redundancy, it is worth a look at the VPS plans, which range from the entry-level VPS Einfach to the high-performance VPS Ultra. Smaller web projects with lower availability requirements will find a suitable, cheaper alternative in the Webspace plans.
If you want to dig deeper into the legal and technical fundamentals, the following sources are a good starting point:
The guide to contract design for cloud computing provides a structured checklist for SLA clauses. The BSI material on high availability explains the technical basics behind availability and data center requirements in detail. For concrete calculation examples, the SLA uptime calculator is useful, as it converts percentages directly into time values. If you are interested in the data protection side of German hosting contracts, you will also find the data protection hosting checklist with practical advice for operators and data protection officers.
Uptime is the share of time in which an IT service is actually reachable and working, usually expressed as a percentage over a given period. Its counterpart is downtime, meaning the time in which the service is not available from the user's point of view.
Uptime is a pure measurement, that is, the actual or expected availability of a service. An uptime guarantee turns it into a contractual promise with legal consequences if the value is not met, provided the contract is worded accordingly.
At 99.9% availability, the calculated downtime budget is about 43 minutes and 50 seconds per month, based on an average month of 30.44 days. At 99.99%, this drops to roughly 4 minutes and 22 seconds per month.
Depending on the contract, service credits, price reduction, damages, or a special right of termination may be available. Complete documentation through monitoring logs and ticket histories is important, because without evidence even justified claims are hard to enforce.
For private or community game servers, a lower level is often sufficient in practice than for business-critical applications, while production services usually aim for values between 99.9 and 99.99%. With game servers, stability under load matters more than the raw percentage, meaning low latency and low packet loss during peak hours.