logging-observability

Structured logging uses JSON format with essential and recommended fields to make logs machine-parseable. This Skill defines the required fields and provides best practices for structuring log data so that aggregation platforms like ELK or Datadog can index and search efficiently.

Tracing requests across microservices requires generating and propagating correlation IDs through service calls. This Skill explains how to attach correlation IDs to logs and pairs them with OpenTelemetry distributed tracing to provide full visibility into request flows across distributed systems.

Prometheus client libraries collect custom metrics using Counter, Gauge, and Histogram types. This Skill provides implementation examples for each metric type, showing how to instrument application code to expose metrics endpoints for scraping and aggregation.

Actionable alerting requires clear principles for when to trigger notifications and includes a YAML template for defining alert rules. This Skill outlines alert design principles to ensure alerts are meaningful, non-noisy, and directly linked to system health indicators.

Health check endpoints distinguish between service availability and readiness: /health provides general status, /ready indicates if the service can accept traffic, and /live shows if the process is running. This Skill details when to use each endpoint for robust orchestration.

Comprehensive observability coverage requires checking structured logging, tracing, metrics, alerting, and health endpoints. This Skill provides a detailed observability checklist to verify all system components are monitored and debuggable across the stack.