Radiator Server Documentation — v10.34.2
Table of Contents

Service level objective

These fields configure an explicit service_level_objective block on a backend or backend server. failure_rate is a failure count and window size such as 3/5. The backoff periods control recovery probes after the failure threshold is reached, and recovery_probe_count sets the number of successful probes needed to resume normal requests.

An omitted field in an explicit block uses the defaults shown in the snippets. See Service Level Objective for failure classification and recovery behavior.