Beyond the uptime guarantee: Key points for UK data centre SLAs

In England and Wales, data centre Service Level Agreements (SLAs) are governed by ordinary principles of contract law. Yet they are frequently misunderstood as performance guarantees, rather than what they truly are: commercial mechanisms for allocating risk between provider and customer.

When a UK colocation provider promises "99.99% uptime" but a one hour outage costs your business £500,000 while the SLA remedy delivers a £10,000 service credit, the limitations of an SLA quickly become apparent. Understanding what SLAs actually protect, and what they do not, is essential when negotiating data centre agreements.

Our view

In our experience, the most common mistake is treating the headline uptime percentage as a guarantee of business continuity. Even where a provider clearly breaches its SLA commitments, the contractual remedy is usually limited to a modest service credit, often capped at one month's fees and expressly stated to be the customer's exclusive remedy. 

These credits rarely reflect the scale of real operational or financial loss as explored in this article on 'UK and Irish businesses'. This gap is increasingly recognised. Many organisations now plan on the assumption that SLAs will be breached at some point and that contractual remedies will be inadequate. That has driven greater focus on technical resilience, business continuity planning, and insurance solutions, including emerging products designed to respond specifically to service level failures. 

According to the National Law Review, specialist SLA insurance policies are now gaining traction as data centre projects scale. The key lesson is to treat the SLA as one element of a broader, layered risk strategy, not as a safety net in its own right.

1. SLAs are risk allocation tools, not insurance

SLAs define service commitments and specify the consequences if those commitments are missed. Under English law, they are enforceable contractual promises, but they are not intended to make customers whole. Service credits are a pricing adjustment mechanism, not compensation for loss. They should be viewed as a baseline allocation of risk, supplemented by technical design, operational planning, and insurance.

2. Uptime and power resilience

Headline availability figures (typically 99.9% to 99.99%) are less important than how downtime is defined. In some colocation SLAs, 'downtime' means a complete loss of power to the customer's rack. Failures of a single feed in a dual supply configuration may not count, nor will outages during properly notified maintenance windows. Customers should ensure that SLA commitments align with the facility's actual power design (for example N+1 or 2N resilience). Disputes frequently arise where marketing claims overstate what the infrastructure can genuinely deliver.

3. Environmental conditions

Environmental SLAs commonly commit to temperature and humidity ranges at rack intake, often aligned with widely accepted industry standards. The detail matters for factors such as where measurements are taken, how long excursions are permitted before a breach occurs, and whether customers have access to real time monitoring data or alerts. Vague drafting in this area can undermine otherwise robust commitments.

4. Connectivity and network services

Network availability is sometimes governed by a separate SLA, with its own availability targets and remedies. These SLAs typically include clear demarcation points defining where the provider's responsibility ends. Cross connect provisioning times are also a crucial factor. Customers should understand these boundaries clearly and design redundancy accordingly..

5. Security and access

SLAs and associated schedules usually describe physical security controls, access arrangements, and alignment with recognised standards. Customers should confirm that access rights, security measures, and audit provisions are sufficient to meet both operational needs and regulatory obligations.

6. Maintenance, incident response, and support

Planned maintenance is almost always excluded from SLA calculations, provided notice requirements are met. Incident response times for critical events are commonly short, but a response does not guarantee a resolution. Remote hands services should be clearly defined, including availability, scope of assistance, and which tasks are included in the core service versus those charged separately.

7. Remedies, liability, and termination rights

UK data centre SLAs deliberately limit remedies. Service credits are typically capped, and overall liability is commonly limited to a defined multiple of fees, with indirect or consequential losses excluded. While such clauses must meet the statutory requirement of reasonableness under English law, customers should assume that recovery for major losses will be constrained. For this reason, termination rights for repeated or chronic SLA failures are often more valuable than marginal increases in service credit percentages.

8. Managing risk beyond the SLA

Because contractual remedies are limited, organisations should plan on the basis that an SLA breach will not cover the true cost of an outage. Practical risk mitigation includes multi site deployments, tested failover arrangements, robust backup strategies, business interruption insurance, and consideration of newer insurance products aimed at service level failures.

9. Exit and decommissioning

Well drafted SLAs and associated terms address exit clearly: notice periods, decommissioning timelines, equipment removal, and responsibilities for cabling and data destruction. Clear exit provisions reduce operational risk and avoid disputes at the end of the relationship.

Practical takeaways

• Know the limits: SLAs provide limited, predefined remedies that rarely reflect the scale of real losses in the event of a major problem.
• Negotiate clarity: Ensure definitions, exclusions, and commitments align with actual capabilities.
• Plan beyond the contract: Design resilience with the knowledge of the SLA’s limitations.
• Use expertise: For complex arrangements, specialist advice from our technology and digital team can materially improve risk allocation and outcomes.