tsuga-investigate-service-health

Service health investigation combines registry checks, error aggregations, request-rate trends, and p95 latency analysis to determine if a named service is degraded. It verifies the service exists, collects owner context, and aggregates rolling error counters over a configurable time window.

Yes, multi-signal investigation computes error counts, request rate trends, and p95 latency by operation when traces exist. It identifies structural error clusters and team-scoped error pattern increases, then synthesizes findings using explicit evidence rules.

Yes, Tsuga CLI access is required to investigate service health. The workflow uses aggregation scalar and timeseries outputs from the CLI to satisfy evidence rules, ensuring safe synthesis that avoids ungrounded causal claims during incident response.

Use service-focused triage to verify the service in the registry and collect source coverage plus rolling error counters for the last 24 hours. This applies to active incidents, first-response triage, and recurring health checks across optional environments.

Trace-log correlation enables p95 latency computation by operation when traces exist, allowing the investigation to identify structural error clusters and team-scoped error pattern increases. This multi-signal approach synthesizes findings using explicit evidence rules.

Yes, service health investigation applies to recurring health checks for a specific service across optional environments and time windows. It combines registry checks, error aggregations, and request-rate trends to monitor ongoing service health.