Status: Canonical Date: 2026-04-26 Owners: OSHUN Reliability, OSHUN Platform
Engineering Tracks: V1-OBS-010 Spec source:
libs/oshun/analytics/src/dashboards-service-health.ts
This manifest defines the per-service health dashboard for every Oshun-critical service and shared dependency.
Targets#
Every entry in OSHUN_V1_TRACING_SUBSYSTEMS (shells, BFFs, Sophia, Iris,
Psyche, Lilith, Isis) plus dependency_postgres, dependency_redis,
dependency_object_storage, dependency_search_index, dependency_queue.
Metrics (per dashboard)#
The four golden signals plus SLO compliance: availability, request_volume,
error_rate, latency_p50 / p90 / p99, saturation_cpu / memory, and
slo_compliance.
SLO and latency targets#
Each dashboard declares an SLO percentage and a p99 latency target. Tests lock SLO ≥ 99% and a positive latency target on every entry.
Panels#
Five canonical panels per dashboard: availability+SLO, error rate, latency distribution, request volume, saturation. Tests lock unique IDs and metric-reference integrity.
Governance#
Adding a service-health target requires updates to the manifest, tests, and this document. Latency / SLO targets are revisited each release.