DevOps Dictionary

Service Level Agreement (SLA) is a formal contract between a service provider and a customer that sets measurable expectations for service performance and support, such as availability (uptime), latency, support response and resolution times, maintenance windows, and data durability. It addresses the problem of vague reliability promises by turning them into specific targets, defining how those targets are measured (metrics, monitoring sources, and reporting periods), and stating what happens if they are missed, often including escalation steps and remedies like service credits. At a high level, an SLA works by selecting key indicators, setting thresholds and exclusions (for example, planned maintenance), and establishing incident handling and communication rules so performance can be tracked consistently.

With a Service Level Agreement (SLA), teams can plan capacity, on-call coverage, and downstream dependencies around clear guarantees; without it, expectations become subjective, disputes are harder to resolve, and outages can translate into unmanaged operational and business risk. This gap exists because shared, verifiable targets align priorities and create accountability for reliability work.