observability-patterns

Observability requires three pillars working together: metrics measure system performance (RED and USE patterns), logs provide detailed events with correlation IDs for tracing requests, and traces show distributed request flows. This Skill covers patterns for all three using Prometheus, Grafana, Datadog, New Relic, and ELK stacks.

RED metrics track Rate, Errors, and Duration for user-facing services; USE metrics track Utilization, Saturation, and Errors for resources. RED applies to API endpoints and services; USE applies to infrastructure. This Skill provides patterns for implementing both across your stack.

Structured logging with correlation IDs lets you trace a single request across multiple services and logs. Each request gets a unique ID passed through all downstream calls, enabling you to aggregate logs and reconstruct the full request path for debugging and root cause analysis.

Yes. OpenTelemetry is the standard for distributed tracing and works across Prometheus, Grafana, Datadog, New Relic, and ELK. This Skill includes patterns for instrumenting services with OpenTelemetry and building dashboards to visualize traces.

SLOs (Service Level Objectives) set reliability targets; SLIs (Service Level Indicators) measure actual performance against those targets; error budgets define acceptable downtime. This Skill covers SLI/SLO definitions, tracking, alerting rules, and post-mortems for incident response.

Production observability requires monitoring stack configuration, alerting rules and channels, runbooks for incident response, health checks, dashboards for visibility, and post-mortem processes. This Skill provides end-to-end patterns for all components across modern stacks.