sentry-instrumentation

What problem does it solve?

This Skill prevents metric sprawl, tag cardinality explosions, and silent metric drift by forcing a governed contract for how Sentry system metrics are named, tagged, emitted, and versioned.

Core Features & Use Cases

• Governed metric grammar: consistent MetricDef schemas, constructors, naming rules, and lifecycle/versioning guarantees across codebases.
• Safe tagging & failure taxonomy: closed sets for tag values and bounded failure classes using classify(), never raw exception strings.
• Correct emission patterns by surface: standardized instrumentation for HTTP routes, external API calls, workflow steps, retry loops, and fallback paths.
• Cost and loop safety: sampling, rate limiting, and loop aggregation enforcement to keep Sentry bills predictable.
• CI and test enforcement: AST-based checks plus pytest gates block invalid metrics before merge.

Quick Start

Ask your AI coding agent to instrument the external API client calls using the skill’s InstrumentedHttpClient pattern and standard triad (counter + latency + failure).