observability-standards

Use the namespace_subsystem_name_unit convention with appropriate metric type suffixes. This Skill provides naming standards and instrumentation guidelines to ensure consistent metrics implementation across API endpoints, database queries, background jobs, and external API calls, eliminating fragmented observability practices.

Instrument using Counter, Gauge, Histogram, and Summary metric types with low-cardinality labels, operation-based metrics for each component, and tracing context propagation. This Skill guides implementation across all service layers with a standardized review checklist to ensure alert-ready metrics.

Propagate correlation IDs and distributed trace context through your instrumentation. This Skill integrates tracing guidelines with metrics standards so you can correlate latency, errors, and performance across services for comprehensive observability.

Collect operation-based metrics tailored to each component type: latency histograms and request counters for APIs, hit/miss rates for caches, and execution time and status for background jobs. This Skill provides domain-specific instrumentation patterns for each use case.

Yes. Metrics are named and structured to be alert-ready, with guidance on cardinality and labeling that prevent alert explosion. This Skill includes alerting considerations so your metrics directly support reliable alert definitions.

Counter tracks cumulative increments, Gauge measures point-in-time values, Histogram records latency distributions, and Summary provides quantile calculations. This Skill explains when to use each type and how to apply them correctly across your Go services.