Observability
Overview
Define how the system tells you it's broken — before it's broken. Output is .forge/observability.md: structured logging conventions, correlation ID flow, the metrics taxonomy (golden signals per service, USE for resources), trace sampling policy, SLO + alert thresholds (page-worthy vs ticket-worthy), dashboard layouts, log retention, and PII redaction rules. Pairs with error-handling-and-resilience (errors classified there get observed here) and `incident-resp
[Description truncada. Veja o README completo no GitHub.]