Setup Observability and Monitoring Systems

Monitoring and Observability Setup is a skill for implementing comprehensive monitoring solutions across the three pillars of observability - metrics, logs, and traces - setting up metrics collection, distributed tracing, and log aggregation, and building dashboards and alerting that give full visibility into system health and performance. It works by clarifying goals, constraints, and required inputs, applying relevant monitoring best practices, and providing actionable steps with verification; when a deeper worked example is needed it opens a separate resources/implementation-playbook.md for detailed patterns.

Its output follows a fixed structure: an infrastructure assessment of current monitoring capabilities, a complete monitoring architecture and stack design, a step-by-step implementation plan, a comprehensive metrics catalog with metric definitions, ready-to-use Grafana dashboard templates, detailed alert-response runbooks, service level objective (SLO) definitions with error budgets, and a service instrumentation integration guide. The stated goal throughout is a monitoring system that surfaces actionable insight, reduces mean time to resolution (MTTR), and enables proactive issue detection rather than purely reactive firefighting.

Use it when working on monitoring and observability setup tasks, or when guidance, best practices, or a checklist for standing up monitoring infrastructure is needed. It is not for tasks unrelated to monitoring and observability setup, or for a different domain or tool outside this scope - and it explicitly stops to ask for clarification if required inputs, permissions, safety boundaries, or success criteria are missing, rather than guessing at them.

Input is the current monitoring setup (if any) plus the specific requirements to address, supplied as free-form arguments. Output is the eight-part deliverable above - assessment, architecture, implementation plan, metrics catalog, Grafana dashboards, alert runbooks, SLOs and error budgets, and an instrumentation guide - not a substitute for environment-specific validation, testing, or expert review.

SREs, platform engineers, and developers who need to design or improve a monitoring and observability stack around metrics, logs, and traces, and want a structured deliverable covering dashboards, alerting, and SLOs rather than ad hoc tool setup.