Observability Coverage Auditor

The problem

Engineering teams often realize their observability is insufficient only after a major production incident occurs. Finding blind spots in logs, metrics, and traces across distributed services is manual, error-prone, and reactive.

What it does

• Audits critical production paths to ensure failures are detectable and diagnosable.
• Inventories telemetry across logs, metrics, traces, and correlation IDs to find coverage gaps.
• Evaluates queue depth, job execution, and external API integration visibility.
• Reviews alert quality and SLO signals to ensure on-call engineers receive actionable data.
• Generates a prioritized remediation plan and a 100-point observability coverage score.

Why this beats prompting it yourself

Writing a prompt to check "if my logging is good" misses the complex interplay between async jobs, distributed tracing, and business logic outcomes. This skill uses a structured 24-step rubric to verify correlation across the entire stack, ensuring you don't just have more data, but better answers during an incident.

Use cases

• Pre-release production readiness reviews for new microservices.
• Post-mortem analysis to identify why an incident wasn't detected or was hard to debug.
• Auditing external integration points and webhook pipelines for silent failures.
• Validating that PII is not leaking into logs or high-cardinality metric labels.

Known limitations

Cannot access live production systems or change alert rules. If runtime telemetry is not provided, results are limited to static design reviews.