Disclaimer: This article provides a practical procurement framework and does not constitute legal advice.
A dedicated server service level agreement (SLA) should define what the hosting provider guarantees, how service failures are measured, and what compensation you can claim. However, the headline uptime percentage reveals only part of the agreement.
Before ordering a dedicated server, you also need to examine hardware replacement, support boundaries, maintenance exclusions, backups, security responsibilities, and cancellation terms. These policies determine what the provider must do when something goes wrong.
This guide explains what to check before signing. It complements our best dedicated hosting providers guide, which compares server configurations, performance, pricing, and provider suitability.
What Documents Form Your Dedicated Hosting Agreement?
Your dedicated hosting agreement may include several documents. The SLA is only one of them.
Review the following before approving an order:
- Order form: Specifies the server configuration, billing term, setup fees, and purchased services.
- Master service agreement: Defines payment, liability, termination, and dispute terms.
- Service level agreement: Defines availability commitments, exclusions, and service credits.
- Acceptable Use Policy: Restricts prohibited content, applications, traffic, and server activities.
- Managed services policy: Separates provider responsibilities from customer responsibilities.
- Backup or disaster recovery policy: Defines backup frequency, retention, restoration, and recovery coverage.
- Data Processing Agreement or BAA: Covers regulated data where GDPR, HIPAA, or similar requirements apply.
Check which document takes precedence when the terms conflict. A negotiated order form should clearly override standard website terms where applicable.
Save dated copies of every document. Your providers can update online policies after you purchase the server.
What Does the Uptime SLA Actually Measure?
A dedicated server SLA may measure network availability without guaranteeing that your website or application will remain online. For example, a provider may consider its service available when your server can transmit and receive data. The guarantee may not cover a failed operating system, corrupted database, control panel problem, application error, or overloaded server.
Check these five parts of the uptime clause:
- Covered components: Confirm whether the SLA covers network connectivity, power, cooling, physical hardware, operating systems, and applications.
- Measurement source: Determine whether the provider accepts external monitoring or relies solely on its internal logs.
- Measurement period: Most providers calculate availability across a calendar month rather than per incident.
- Outage timing: Establish whether downtime begins when monitoring detects the failure or when you open a support ticket.
- Excluded incidents: Review exclusions for scheduled maintenance, emergency maintenance, DDoS attacks, customer errors, third-party services, and force majeure events.
Also check whether severe packet loss or latency counts as downtime. A server can technically remain reachable while delivering unusable application performance.
Example: Atlantic.Net Measures Downtime From the Support Ticket
Atlantic.Net guarantees 100% network availability for dedicated servers, excluding scheduled maintenance. However, its policy measures network downtime from the time a trouble ticket is opened and requires the failure to be recorded in its ticket system.
The guarantee covers routers, switches, cabling, and network connectivity. It does not cover software or services running on the server. This wording shows why customers need independent monitoring and a clear incident-reporting procedure. See screenshot below for your reference; or you can review the full terms in their Dedicated Server Service Level Guarantee.
How Are SLA Credits Calculated and Claimed?
SLA credits compensate you for qualifying downtime, but they rarely compensate for lost sales, employee time, or reputational damage.
A typical credit calculation uses the affected server’s monthly fee. If the server costs $500 per month and the SLA awards a 10% credit, you receive $50 toward a future invoice. You do not necessarily receive $50 in cash.
Before signing, confirm:
- The credit percentage for each outage threshold
- Whether downtime is rounded into fixed intervals
- The maximum credit allowed each month
- Whether credits apply to future invoices only
- The deadline for submitting a claim
- The evidence required to support the claim
- Whether overdue accounts lose credit eligibility
- Whether service credits are your only contractual remedy
A short claim window can make a generous credit policy difficult to use. Assign someone to retain monitoring logs, ticket numbers, timestamps, and provider responses after every incident.
What Support and Hardware Commitments Should You Verify?
Support response, problem resolution, and hardware replacement are different commitments.
An initial response confirms that the provider has received your ticket. It does not guarantee that an engineer has diagnosed the problem or restored the server. Ask the provider to define response and restoration targets for each incident severity.
Hardware replacement clauses also require close attention. The replacement timer might begin only after the provider confirms that a component has failed. Time spent diagnosing the server may fall outside the guarantee.
Check whether the hardware SLA excludes:
- Reinstalling the operating system
- Reconfiguring applications
- Restoring files and databases
- Cloning a failed drive
- Rebuilding a RAID array
- Troubleshooting customer-installed software
Example: LiquidWeb Replaces Faulty Hardware Within 30 Minutes
Liquid Web, for example, states that faulty dedicated server hardware will usually be replaced within 30 minutes after the problem is identified. Its guarantee excludes software maintenance, account restoration, operating system reloads, application configuration, and RAID rebuilding.
Liquid Web also lists targets of under one minute for chat responses and under one hour for help-desk tickets. However, the provider explicitly states that these response times are not guaranteed under its SLA. The Liquid Web Service Level Agreement demonstrates the difference between a published service target and an enforceable commitment.
What Should Your Dedicated Server Procurement Checklist Include?
A procurement checklist should cover the complete service, not only the physical server.
| Area | What You Should Confirm |
|---|---|
| Server configuration | Exact CPU model, RAM, drive type, RAID level, controller, spare-part availability, and component replacement terms |
| Network | Port speed, monthly bandwidth, overage charges, routing, IP allocation, DDoS protection, and traffic-filtering limits |
| Server management | Operating system updates, control panel support, security patching, service monitoring, and application troubleshooting |
| Backups | Backup frequency, retention period, storage location, encryption, restoration procedure, and restoration fees |
| Recovery | Recovery Point Objective, Recovery Time Objective, failover options, and disaster-recovery testing |
| Security | Access controls, vulnerability management, logging, incident notification, and responsibility for compromised accounts |
| Maintenance | Notice period, scheduled windows, emergency maintenance rules, and excluded downtime |
| Support | Available channels, severity levels, response targets, escalation contacts, and remote-hands charges |
| Billing | Setup fees, renewal rates, software licences, bandwidth charges, management add-ons, and price-change notice |
| Exit terms | Cancellation notice, early termination fees, data-export period, server access after cancellation, and secure data deletion |
Important promises should appear in the signed order or contract. A sales email can clarify the offer, but it may not override the provider’s standard terms.
What Should You Do Before Signing?
Review the final agreement against the workload you intend to host. A general 100% network SLA may be insufficient for an eCommerce store, healthcare application, SaaS platform, or database that needs defined recovery times.
Before approving the purchase:
- Mark every undefined or ambiguous SLA term.
- Calculate the real credit for a realistic outage.
- Confirm who manages patching, monitoring, backups, and restoration.
- Test the provider’s support and escalation channels.
- Request written exceptions for any negotiated promises.
- Save the signed order and all incorporated policies.
Once these requirements are documented, use our best dedicated hosting providers guide to compare suitable plans and vendors.
