Skip to main content

Split Cause

Graph-based root cause analysis with Bayesian inference. Model system dependencies, gather evidence, and let the inference engine guide your investigation.

Features

  • Visual graph modeling of system dependencies
  • Bayesian probability updates based on evidence
  • Hypothesis testing with confidence tracking
  • Suspect slice computation (Reach ∩ RevReach)
  • Conflict resolution for mutually exclusive hypotheses
  • Snapshots and versioning
  • Share links with scoped access control

How It Works

  1. 1.Create nodes for system components and hypotheses
  2. 2.Draw edges to model dependencies and flows
  3. 3.Mark components as OK/FAILED based on observations
  4. 4.Attach evidence to nodes (logs, metrics, test results)
  5. 5.Run inference to update hypothesis probabilities
  6. 6.Identify likely root causes based on posteriors

Recent Investigations

No recent investigations yet.

Create your first investigation to get started!