An SLO is the reliability target you commit to, and it is what turns 'be reliable' into a number you can manage against.
A Service Level Objective (SLO) is a target for a reliability metric over a time window, for example 99.9% of requests succeeding over 30 days. It sets the agreed bar for how reliable a service should be.
These three fit together. The SLI is the measured indicator (for example, the percentage of requests served under 300ms). The SLO is the target on that indicator (99.9% over 30 days). The SLA is the external contract with consequences. The SLO is the internal goal that drives engineering decisions.
An SLO makes reliability a deliberate trade-off rather than an absolute. Perfect reliability is impossibly expensive, so you set a target that balances user expectations against cost and velocity, and you manage to it. The gap between the SLO and 100% is the error budget you are allowed to spend.
Ops Singularity ties resolution to reliability targets: Sentinel AI can prioritise incidents by their impact on SLOs, acting first where an error budget is most at risk.
An SLO is an internal reliability target; an SLA is an external contract, often with financial consequences, usually set looser than the SLO.
Base it on user expectations and the cost of reliability; common targets are 99.9% or 99.95%, chosen so the error budget leaves room to ship changes.
Ops Singularity turns open telemetry into autonomous, governed resolution. See it on your own stack.