Skip to main content

konditional-observability

konditional-observability adds shadow evaluation helpers and exposes observability hooks.

When to Use This Module​

You should use konditional-observability when you need to:

  • Compare two configurations side-by-side with shadow evaluation to detect mismatches
  • Add instrumentation without coupling to specific logging or metrics vendors
  • Get explainable evaluation results for debugging and auditing
  • Track evaluation performance and decision paths in production

What You Get​

  • Shadow evaluation: Compare baseline and candidate configurations safely
  • Dependency-free hooks: KonditionalLogger and MetricsCollector interfaces
  • Evaluation explanations: EvaluationResult with decision traces and bucketing info
  • Mismatch detection: Callbacks when shadow evaluation reveals differences

Alternatives​

Without this module, you would need to:

  • Build custom shadow evaluation logic with careful isolation to avoid affecting production
  • Tightly couple feature flag evaluation to specific logging/metrics libraries (vendor lock-in)
  • Implement your own evaluation tracing and debugging utilities from scratch

Installation​

dependencies {
  implementation("io.amichne:konditional-observability:VERSION")
}

Guarantees​

  • Guarantee: Shadow evaluation does not alter production behavior.

  • Mechanism: The baseline registry is returned; the candidate registry is evaluated for comparison only.

  • Boundary: Mismatch callbacks run inline on the evaluation thread and must stay lightweight.

Next steps​