What is an SLA (Service Level Agreement)?
An SLA (service level agreement) is an agreement between you and a provider that defines what level of service you can expect and how it is measured. In an IT context, an SLA is typically part of the support or operations contract and describes concrete numbers for response times, resolution times and uptime. Large cloud providers such as Microsoft, for example, publish SLAs with guaranteed uptime for their services.
The point is simple: Without an SLA, "fast support" and "stable operations" are just promises. With an SLA, they are commitments you can hold the provider to.
What does SLA mean?
SLA is an abbreviation of Service Level Agreement. In essence, it is the part of a contract that makes service measurable. Note that SLA in an IT context has nothing to do with other meanings of the abbreviation; if you are looking for a definition within IT, the service level agreement is the one you want.
What should an SLA contain?
A good SLA contract should as a minimum cover:
- Response times and resolution times: how quickly does the provider react, and how quickly is the problem solved, broken down by priority.
- Uptime: for example 99.9% measured per month, and how it is calculated.
- Maintenance windows and exceptions: planned maintenance, force majeure and faults outside the provider's responsibility.
- Escalation paths: who is contacted if a case is not resolved within the agreed time, and when management is involved.
- Penalties for breaches: typically a penalty or compensation. An SLA without consequences is really only a statement of intent.
- Reporting: how and how often you receive documentation that the levels are being met.
SLA example: P1-P4 response times
Most SLAs in IT prioritise cases by business impact. A typical model looks like this:
| Priority | Example | Response time |
|---|---|---|
| P1 (critical) | The entire business is down | 15-30 minutes |
| P2 (high) | A department or a core system is affected | 1-2 hours |
| P3 (normal) | A single user is affected, work can continue | 4-8 hours |
| P4 (low) | Requests, changes, questions | 1-3 working days |
The numbers vary from agreement to agreement. What matters is that the priorities are defined up front, so a crashed server does not end up in the same queue as a printer question.
SLA vs. operations agreement
An operations agreement describes what the provider does: monitoring, backup, updates, support. The SLA describes how well it is done: response times, uptime and consequences for breaches. They belong together: An operations agreement without an SLA lacks measuring points, and an SLA without an operations agreement lacks substance. In the same vein, you should also know your requirements for recovery after an outage, see disaster recovery.
How MI Support IT can help
MI Support IT works with clear, measurable SLA terms as a standard part of our agreements, including a 24/7 service desk with defined response times across all priorities. See our pricing and agreement types, or contact us if you would like your current SLA reviewed with fresh eyes.