The internal target for an SLI, e.g. '99.9% of requests under 200ms over 30 days', the contract a team holds itself to.
A Service Level Objective (SLO) is a target value or range for a service-level indicator (SLI), set by the engineering team that owns the service. 'p95 checkout latency under 500ms over 30 days', '99.9% availability over 30 days'. The SLO is more strict than the customer-facing SLA so the team has buffer for occasional misses without breaching the external contract. Pair every SLO with an error budget to convert reliability into a tradeable currency.
SLOs convert vague aspirations ('the service should be reliable') into concrete numbers that can be measured, alerted on, and traded against feature work. Without SLOs, reliability investment always loses to feature pressure; with them, the math forces a balance.
See the part of the platform that handles slo (service level objective) in production.