# Decide the telemetry, SLO, retention, and alerting control plane

Type: grilling
Status: open

## Question

For the accepted initial single-host service, decide the log/metric/trace
collection and storage path, instance roles, release and correlation identity,
redaction/access/retention policy, initial service-level indicators and
objectives, error budgets, alert thresholds, on-call destinations, escalation
ownership, and telemetry-blind detection. Prefer the smallest operable control
plane, but record every domain-visible availability or retry assumption in the
appropriate ADR/spec rather than hiding it in dashboard configuration.