Skip to main content

Guides

What Is an SLA in IT Maintenance?

Response time, resolution time, severity levels, uptime targets and service credits explained, so you can read a maintenance SLA with confidence.

Published
Updated
Reading time
7 min read
Ticket resolvedEngineer assigned

When you hand responsibility for your website, applications or servers to a maintenance provider, the service level agreement (SLA) is where expectations become specific. It sets out what the provider will do, how quickly, during which hours, and what happens if they fall short. A good SLA protects both sides: you know what to expect, and the provider knows exactly what they have committed to.

The trouble is that SLAs are full of terms that sound similar but mean very different things. "One-hour response" is not the same as "fixed within one hour". A "99.9% uptime target" may exclude more than you assume. And a service credit is rarely compensation for the business impact of an outage.

This guide explains the main components of an IT maintenance SLA in plain English, using illustrative examples. The figures shown are examples to explain the concepts, not commitments from SupportSafe or any particular provider; our own service levels are set out in each plan and agreement.

What an SLA is – and is not

An SLA is a defined set of measurable service commitments, usually forming part of a wider contract or terms of service. It typically covers:

  • The services in scope and the assets they apply to.
  • Support hours and how to raise requests.
  • How issues are classified by severity.
  • Target response and resolution times for each severity.
  • Availability or uptime targets, where the provider controls the infrastructure.
  • Exclusions and dependencies.
  • Reporting and review arrangements.
  • Remedies, such as service credits, if targets are missed.

An SLA is not a guarantee that nothing will go wrong. It is a commitment about how the provider will respond when something does, and how service quality will be measured.

Response time versus resolution time

This is the most misunderstood part of any SLA.

  • Response time is how long it takes for a qualified person to acknowledge the issue and begin working on it. An automated "we have received your ticket" email usually does not count; a good SLA defines response as meaningful human engagement.
  • Resolution time is how long it takes to restore the service or fix the problem. Many SLAs distinguish between a workaround (service restored, root cause still being addressed) and a permanent fix.

Response times are much easier for a provider to commit to than resolution times, because resolution can depend on factors outside their control – a hosting provider outage, a third-party API failure, or a vulnerability with no available patch. That is why many maintenance SLAs set firm response targets and "target" or "best-effort" resolution times, sometimes with commitments to provide regular updates until resolution.

When comparing providers, look carefully at which of the two is being promised.

Severity levels

Severity levels allow the provider to prioritise work according to business impact. Definitions should be clear enough that both sides would classify the same incident in the same way. The table below shows an illustrative structure.

Severity (example)Typical definitionExample incidentsExample response targetExample update frequency
P1 – CriticalService down or severely impaired for all users; security incident in progressWebsite offline, checkout failing, confirmed compromiseWithin 1 hour, 24/7Every hour until resolved
P2 – HighMajor function impaired; no reasonable workaroundPayment method failing, contact forms not delivering, severe slowdownWithin 4 business hoursTwice daily
P3 – MediumPartial impairment with a workaround availableLayout issue on one template, a non-critical integration delayedWithin 1 business dayAs progress is made
P4 – LowMinor issue, question or cosmetic changeTypo, small content update, general adviceWithin 2–3 business daysOn completion

Again, these figures are purely illustrative. Real targets vary by plan, provider and the hours of cover purchased.

Who decides the severity?

A good SLA explains how severity is assigned – typically by the client when raising the issue, confirmed or adjusted by the provider based on agreed definitions – and how disagreements are resolved. It should also explain how to raise a critical issue out of hours, since email may not be monitored overnight.

Support hours and coverage

Response targets only make sense alongside support hours. "Four-hour response" within business hours means a ticket raised at 5pm on a Friday might not be picked up until Monday morning. Check:

  • Which hours and days count as business hours, and in which time zone.
  • Whether bank holidays are included.
  • Whether 24/7 cover applies to all severities or only critical incidents.
  • How out-of-hours emergencies are raised – a phone line, a dedicated email or an alerting system.

Uptime targets

Where the provider manages hosting or infrastructure, the SLA may include an availability or uptime target, usually expressed as a percentage over a calendar month. It is worth translating percentages into time to understand what they allow. Using a 30-day month as an example:

Example uptime targetApproximate permitted downtime per 30-day month
99%About 7 hours 12 minutes
99.5%About 3 hours 36 minutes
99.9%About 43 minutes
99.95%About 22 minutes

The questions that matter are how uptime is measured (from where, how often and against which URL or service), what counts as "down" (complete failure only, or also severe slowness and errors), and what is excluded. No responsible provider guarantees 100% uptime, and you should be cautious of anyone who does.

Uptime targets also only apply to what the provider controls. If your provider maintains your website but the hosting is with a separate company, the maintenance provider can monitor uptime and respond to outages, but cannot commit to the hosting company's availability. Our website performance monitoring service is designed to provide that visibility.

Exclusions and dependencies

Every SLA has exclusions, and they shape what the headline numbers really mean. Common examples include:

  • Scheduled maintenance windows, normally announced in advance.
  • Third-party failures – hosting providers, DNS, CDNs, payment gateways, email services and external APIs outside the provider's control.
  • Client-caused issues – changes made by the client or another supplier without coordination.
  • Force majeure events.
  • Unsupported software – end-of-life platforms or extensions the client chose to keep against advice.
  • Out-of-scope work – new features, redesigns and large projects, which are usually quoted separately.

Exclusions are normal and reasonable; the key is that they are clear and that you understand them before signing.

Service credits

Service credits are the most common remedy when an SLA target is missed. Typically, the client receives a credit against future fees, often a percentage of the monthly fee scaled by how far the target was missed, subject to a cap.

Important points to understand:

  • Credits usually need to be claimed within a set period; they are not always applied automatically.
  • They are normally capped, often at a proportion of the monthly fee.
  • They are usually the sole remedy for missed service levels, which limits other claims.
  • They rarely reflect the true business cost of an outage.

That last point is why the provider's processes – monitoring, staging-first updates, tested backups, clear escalation – matter more than the credit schedule. A credit is a signal of accountability, not insurance.

Reporting and review

An SLA without measurement is just a promise. Look for regular reporting on tickets raised, severity, response and resolution performance, uptime where applicable, and work performed. Periodic service reviews give both sides a chance to discuss trends, adjust priorities and agree improvements. Our approach includes monthly maintenance reports and quarterly reviews.

Questions to ask before you sign

  1. Are response and resolution targets clearly distinguished?
  2. Are severity definitions specific enough to apply consistently?
  3. What are the support hours, and how are emergencies raised out of hours?
  4. How is uptime measured, and what is excluded?
  5. How are service credits calculated, claimed and capped?
  6. How often will performance be reported, and in what format?
  7. What happens at the end of the contract – handover of credentials, documentation and backups?

Our guide on how to choose a website maintenance company covers further questions worth asking.

Conclusion

A clear SLA turns a maintenance relationship into something measurable. Understanding the difference between response and resolution, how severity is assigned, what uptime figures really mean and how exclusions and credits work puts you in a far better position to compare providers and hold them to account.

Each of our maintenance plans sets out its support hours and service level targets clearly, and Enterprise agreements can include custom SLAs. If you would like to discuss the service levels your sites or applications need, get a quote and we will recommend an appropriate level of cover.

Frequently asked questions

What is a typical SLA response time for website maintenance?
It varies widely by provider, plan and support hours. Critical incidents usually have the fastest response targets, often with out-of-hours cover on higher plans, while low-priority requests may have targets measured in business days.
What is the difference between an SLA and a contract?
The contract sets out the overall legal and commercial relationship. The SLA is the part, or schedule, that defines measurable service levels such as response times, severity levels and uptime targets.
Can a maintenance provider guarantee 100% uptime?
No responsible provider guarantees 100% uptime. Hardware, networks, third-party services and software can all fail. A good SLA sets realistic targets, defines how they are measured and explains how incidents are handled.
Are service credits paid automatically?
Not always. Many SLAs require the client to claim credits within a defined period. Check the claim process, the calculation method and any cap before you sign.

SupportSafe Engineering Team

Maintenance, infrastructure and security engineers

Articles are written and technically reviewed by the SupportSafe engineers who carry out website, application and server maintenance for our clients. Each article is reviewed for accuracy before publication and updated when practices change.

Let's take responsibility for your technology.

Tell us what you run and what keeps you up at night. We'll assess it and recommend a maintenance plan that fits – with no obligation.