observability-review

You can review code changes for observability gaps by analyzing pull requests or file paths to evaluate span design, context propagation, and metric instrument choice. This process identifies telemetry issues before they reach production.

Span lifecycle correctness ensures OpenTelemetry spans start and end properly without leaking context. Validating span lifecycles prevents broken distributed traces and misleading telemetry data in production environments.

Checking metric cardinality risks involves scanning code diffs for high-cardinality attributes like user IDs in metric labels. Evaluating cardinality discipline prevents expensive telemetry storage and pipeline degradation.

Yes, the review applies Honeycomb-friendly defaults like wide events and trace or log correlation while checking compliance with OpenTelemetry semantic conventions. It provides guidance without enforcing dogma on contested tradeoffs.

Validating SLI and SLO coverage in a pull request requires evaluating metric instrument choices and structured logging patterns against service objectives. This review flags missing service level indicators before code merges.

Finding broken context propagation requires reviewing code diffs to verify OpenTelemetry context flows correctly across service boundaries. Detecting propagation breaks early prevents fragmented traces and lost span relationships.