Skip to content
Kitbay
Skill

observability-and-instrumentation

MITUpdated on Nov 18, 2026118,985GitHub stars

A skill from Addy Osmani's agent-skills collection for adding logging, metrics, tracing and alerts while a feature is being built, so its behaviour in production can be seen and diagnosed.

What the skill teaches

A skill from Addy Osmani's agent-skills collection for adding logging, metrics, tracing and alerts while a feature is being built, so its behaviour in production can be seen and diagnosed. The agent first writes a few questions an on-call engineer would ask, then chooses signals that answer them: structured JSON logs with a correlation ID and no secrets, rate, error and duration metrics with bounded labels and percentiles, OpenTelemetry traces across async boundaries, and alerts on user-facing symptoms, each linked to a short runbook. It ends by triggering failures in staging to confirm the signals appear. Live incident work and profiling are left to sibling skills in the collection.

Capabilities

  • Add structured logs with correlation IDs
  • Instrument rate, error and duration metrics
  • Add OpenTelemetry traces
  • Write symptom-based alerts with runbooks