A service level agreement, or SLA, is the part of a contract that sets out the standards a service provider promises to meet. In IT it usually covers response and fix times, system uptime and support hours, and what happens if the provider falls short of those levels.
What a service level agreement is
An SLA turns vague promises into measurable commitments. Where a contract might say a supplier will "provide a responsive support service", an SLA replaces that with a precise figure - for example, that critical issues will receive a first response within one hour on business days.
SLAs are standard in IT support, cloud services and managed service provider agreements. The ISO/IEC 20000 international standard for IT service management treats agreed service targets as a core requirement precisely because vague commitments lead to disputes.
For a UK SME, an SLA is the document that tells you, before you sign, exactly what you're paying for and what recourse you have when something goes wrong.
What an SLA usually covers
Every SLA is different, but the same categories appear in most IT contracts. Knowing what to look for makes the document much easier to read.
- Response time: how quickly the provider acknowledges a reported issue.
- Resolution or fix time: how long the actual repair or fix is expected to take.
- System uptime or availability: the percentage of time a service or system must be operational, often expressed as 99.9% or similar.
- Support hours: the times of day and days of the week when support is available.
- Priority levels: how issues are categorised - for example, a server outage versus a single user's printing problem.
- Reporting: how and how often the provider demonstrates it is meeting the targets.
- Remedies: what happens when a target is missed, commonly a service credit on your next invoice.
How to read an SLA
The most important distinction is between response time and resolution time. A one-hour response means someone will acknowledge your call within an hour. It does not mean the problem will be fixed within an hour. Both figures matter and a good SLA defines each one separately.
Uptime figures deserve equal scrutiny. A 99.9% uptime target sounds near-perfect, but it still allows roughly eight hours of downtime per year. The NCSC guidance on cloud service reliability notes that availability figures should be read alongside exclusion clauses, which often exempt scheduled maintenance windows from the calculation.
Check the exclusions in any SLA. Events outside the provider's control, such as a third-party outage or a hardware failure the provider didn't cause, are frequently carved out. Understanding what is and isn't covered tells you how much the uptime guarantee is actually worth.
What to check before you sign one
Ask whether the targets are realistic and expressed in plain, measurable terms. Confirm what qualifies as a priority issue and who decides - a supplier that controls its own triage process can downgrade an urgent problem and still technically meet the SLA.
A clear SLA is a marker of good IT governance. It applies whether you use a fully outsourced managed service provider, a shared arrangement like co-managed IT, or a cloud platform. It's also one of the things break-fix support usually lacks entirely - meaning there's no agreed standard to hold anyone to when something fails.
Finally, consider how much your business actually depends on the services covered. The cost of IT downtime is usually higher than businesses expect, so the SLA targets should reflect that reality.
Red Eagle Tech backs its managed IT with clear, realistic SLAs. Our IT Operations service sets out response and resolution targets in plain language, so you know exactly what you're getting and what happens if we fall short.
