Generates hypotheses about likely root causes.

Analyzes patterns across enricher data to suggest investigation paths. Designed to help LLMs and developers focus debugging efforts.

Not enabled by default - run with --analyze=hypothesis.

Output

%{
  findings: [
    %{
      severity: :info,
      message: "Memory grew 5.2x. Likely cause: 'items' list is growing unbounded.",
      events: [],
      metadata: %{
        growth_factor: 5.2,
        investigation_steps: [
          "Check which assigns are growing between events",
          "Look for lists that accumulate without limit",
          "Consider pagination or windowing for large datasets"
        ]
      }
    }
  ],
  stats: %{hypothesis_count: 1}
}