Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

MetaRCA: A Generalizable Root Cause Analysis Framework for Cloud-Native Systems Powered by Meta Causal Knowledge

Shuai Liang, Pengfei Chen, Bozhe Tian, Gou Tan, Maohong Xu, Youjun Qu, Yahui Zhao, Yiduo Shang, Chongkang Tan

TL;DR

MetaRCA shows robust cross-system generalization, maintaining over 80% accuracy across diverse systems, and surpasses the strongest baseline by 29 percentage points in service-level and 48 percentage points in metric-level accuracy.

Abstract

The dynamics and complexity of cloud-native systems present significant challenges for Root Cause Analysis (RCA). While causality-based RCA methods have shown significant progress in recent years, their practical adoption is fundamentally limited by three intertwined challenges: poor scalability against system complexity, brittle generalization across different system topologies, and inadequate integration of domain knowledge. These limitations create a vicious cycle, hindering the development of robust and efficient RCA solutions. This paper introduces MetaRCA, a generalizable RCA framework for cloud-native systems. MetaRCA first constructs a Meta Causal Graph (MCG) offline, a reusable knowledge base defined at the metadata level. To build the MCG, we propose an evidence-driven algorithm that systematically fuses knowledge from Large Language Models (LLMs), historical fault reports, and observability data. When a fault occurs, MetaRCA performs a lightweight online inference by dynamically instantiating the MCG into a localized graph based on the current context, and then leverages real-time data to weight and prune causal links for precise root cause localization. Evaluated on 252 public and 59 production failures, MetaRCA demonstrates state-of-the-art performance. It surpasses the strongest baseline by 29 percentage points in service-level and 48 percentage points in metric-level accuracy. This performance advantage widens as system complexity increases, with its overhead scaling near-linearly. Crucially, MetaRCA shows robust cross-system generalization, maintaining over 80% accuracy across diverse systems.

MetaRCA: A Generalizable Root Cause Analysis Framework for Cloud-Native Systems Powered by Meta Causal Knowledge

TL;DR

MetaRCA shows robust cross-system generalization, maintaining over 80% accuracy across diverse systems, and surpasses the strongest baseline by 29 percentage points in service-level and 48 percentage points in metric-level accuracy.

Abstract

The dynamics and complexity of cloud-native systems present significant challenges for Root Cause Analysis (RCA). While causality-based RCA methods have shown significant progress in recent years, their practical adoption is fundamentally limited by three intertwined challenges: poor scalability against system complexity, brittle generalization across different system topologies, and inadequate integration of domain knowledge. These limitations create a vicious cycle, hindering the development of robust and efficient RCA solutions. This paper introduces MetaRCA, a generalizable RCA framework for cloud-native systems. MetaRCA first constructs a Meta Causal Graph (MCG) offline, a reusable knowledge base defined at the metadata level. To build the MCG, we propose an evidence-driven algorithm that systematically fuses knowledge from Large Language Models (LLMs), historical fault reports, and observability data. When a fault occurs, MetaRCA performs a lightweight online inference by dynamically instantiating the MCG into a localized graph based on the current context, and then leverages real-time data to weight and prune causal links for precise root cause localization. Evaluated on 252 public and 59 production failures, MetaRCA demonstrates state-of-the-art performance. It surpasses the strongest baseline by 29 percentage points in service-level and 48 percentage points in metric-level accuracy. This performance advantage widens as system complexity increases, with its overhead scaling near-linearly. Crucially, MetaRCA shows robust cross-system generalization, maintaining over 80% accuracy across diverse systems.
Paper Structure (36 sections, 16 equations, 8 figures, 4 tables)

This paper contains 36 sections, 16 equations, 8 figures, 4 tables.

Table of Contents

  1. Introduction
  2. Background and Motivation
  3. Background
  4. Root Cause Analysis based on Causal Graph
  5. Fault Reports.
  6. Motivation
  7. Problem Formulation
  8. Detailed Design
  9. Meta Causal Graph Initialization
  10. Metadata Ontology Definition
  11. LLM-based Causal Knowledge Bootstrapping
  12. Meta Causal Graph Evolution
  13. Evidence Extraction
  14. Bayesian Belief Evolution Model
  15. MCG-based Online Root Cause Analysis
  16. ...and 21 more sections

Figures (8)

  • Figure 1: Shifting from system-specific RCA to reusable causal patterns.
  • Figure 2: The framework of MetaRCA.
  • Figure 3: Performance comparison of different RCA methods as the number of microservice nodes increases. (a) Efficiency, measured by execution time. (b) Service-level localization accuracy, measured by AC@3.
  • Figure 4: Distribution of localization accuracy (AC@3) for different RCA methods across seven systems (three open-source and four production). The box plots show the performance distribution at (a) the service level and (b) the metric level.
  • Figure 5: Localization accuracy (AC@3) of different causal graph construction and ranking method combinations. The performance is shown on (a) the RE2-TT dataset and (b) the Production dataset, with results for both service-level (total height) and metric-level (solid portion) accuracy.
  • ...and 3 more figures