What problem does it solve? When something breaks in production, teams often have no instrumentation in place and must guess at the cause. This Skill designs the observability layer up front — what to measure, where to instrument it, and which queries answer operational questions — so incidents are resolved with a query instead of speculation. ## Core Features & Use Cases - Four-layer instrumentation design: Covers product events (signup, checkout, activation), technical metrics (latency, error rate, throughput, saturation), structured JSON logs with request context, and distributed traces across services. - Question-driven instrumentation: Every event, metric, log, or trace must answer a concrete operational question, preventing useless telemetry. - Baseline query authoring: Produces the queries that show system health and the most likely failure modes, feeding into production-readiness reviews and incident response. - Use Case: Before shipping a new checkout flow, invoke this Skill to define the activation events, golden-signal metrics, error-path logs, and trace spans needed, plus the queries to monitor them after launch. ## Quick Start Invoke /observe to design the instrumentation and operational queries for the change you are about to ship.
