What problem does it solve?
Unstructured, missing, or irrelevant telemetry turns production incidents into time-consuming archaeology, leaving teams unable to quickly diagnose failures or verify that new features work as expected in production.
Core Features & Use Cases
- Structured Logging & Correlation IDs: Emit queryable JSON log events with mandatory correlation IDs to reconstruct full request paths across distributed services, with strict rules to avoid logging secrets or PII.
- RED/USE Metrics: Instrument endpoints and dependencies with rate, error, and duration (or utilization, saturation, error) metrics using bounded label sets to avoid cardinality issues, with percentile-based latency tracking instead of averages.
- Distributed Tracing & Alerting: Implement cross-service tracing with OpenTelemetry and set up symptom-based alerts that fire only when users are impacted, linked to runbooks for fast on-call response.
- Use Case: Use this skill when shipping a new payment processing feature to add the right telemetry so on-call engineers can diagnose payment failures in minutes instead of hours.
Quick Start
Use the observability-and-instrumentation skill to add structured logging, RED metrics, and symptom-based alerting to the new user authentication endpoint before it is deployed to production.
