What is an SLA? (Service-Level Agreement)
A Service-Level Agreement (SLA) is a contractual commitment between a service provider and customer defining measurable performance targets such as uptime percentage, response times, and resolution windows, along with remedies (credits or penalties) when the provider misses those targets.
In more detail
SLAs are common in IT services, cloud hosting, managed services, customer support, and staffing. A well-written SLA makes three things explicit: the metric being measured, the target value, and what happens if the target is missed. Cloud providers like AWS publish SLAs such as 99.99% monthly uptime with service credits for breach.
SLAs differ from OLAs (Operational Level Agreements, internal between teams) and UCs (Underpinning Contracts, with sub-vendors). Customers should pay close attention to how uptime is calculated, what is excluded (scheduled maintenance, force majeure), and how credits are claimed.
How it works
- Define the service being covered (scope).
- Choose metrics (uptime, first-response time, resolution time, CSAT).
- Set target values with clear measurement method.
- Specify exclusions (maintenance windows, customer-caused issues).
- Define remedies (service credits, termination rights).
- Schedule regular service reviews and reporting cadence.
Related terms
Mini FAQ
99.9% (about 8.7 hours downtime/year) is common for business SaaS. 99.99% (52 minutes/year) is premium. 99.999% (5 minutes/year) is telecom grade.
Usually yes. Credits are typically capped at a percentage of monthly fees. Material repeated breaches sometimes allow termination.
Yes. Managed staffing SLAs typically cover shortlist delivery time, replacement guarantees, and account-manager response times.