Observability Plan
Plan the logs, metrics, traces, alerts, dashboards, service level objectives, and incident response hooks for a system so its operators can tell within minutes whether users are hurting and why. Starts from user journeys, defines good-event service level indicators and SLO targets with error budget and multi-window burn-rate alert math, maps instrumentation per component, sets log field and metric cardinality rules, designs tiered dashboards and symptom-based paging, adds runbook and escalation hooks, and budgets telemetry volume, retention, and privacy. Use when someone asks how to monitor a system, what to log or alert on, how to set SLOs, why alerts are noisy, how to add tracing, or how to prepare a service for on-call. Do not use for triaging a specific metric anomaly (use anomaly-triage), running a live incident (use coordinate-production-incident), auditing existing code (use codebase-audit), or defining infrastructure resources (use iac-setup).