HERO: Hypothesis-Centered Root-Cause Analysis for Microservice Incidents
Root-cause analysis (RCA) in cloud-native microservice systems is challenging because it requires integrating heterogeneous observability signals and deciding what to inspect next under limited diagnosis budgets. Existing large language model (LLM)-based RCA agents either rely heavily on predefined workflows, whose eff...