What problem does it solve? Production features often ship without telemetry, so when something breaks, engineers cannot tell what happened from the available data. This Skill guides you to instrument code alongside the feature so production behavior is visible, queryable, and diagnosable from day one. ## Core Features & Use Cases - Structured Logging: Enforces JSON log events with stable event names, mandatory correlation IDs, entry-point attribution, and consistent log levels, while blocking secrets and PII from log output. - Metrics and Tracing: Applies RED (Rate, Errors, Duration) and USE (Utilization, Saturation, Errors) patterns with bounded label cardinality, histogram-based latency percentiles, and OpenTelemetry distributed tracing with context propagation. - Alerting and Runbooks: Designs symptom-based alerts with page/ticket severities, thresholds, and linked runbooks, plus a verification step that test-fires alerts and confirms telemetry by inducing failures in staging. - Use Case: You are adding a payment retry flow with an external provider. The Skill walks you through defining on-call questions, emitting a payment_failed structured event, adding a latency histogram for provider calls, tracing the checkout across services, and paging only when the user-facing error rate crosses a threshold. ## Quick Start Use the observability-and-instrumentation skill to add structured logging, RED metrics, tracing, and an alert with a runbook to my new checkout endpoint.
