Root Cause Analysis

Also known as RCA

Root cause analysis asks which underlying fault started a failure, rather than merely identifying the components or signals that look abnormal. It matters in settings such as microservices, networks, industrial systems, and software operations, where faults propagate and symptoms can appear far from their source. Methods range from statistical change detection and causal graph discovery to counterfactual testing, multimodal telemetry analysis, and language-model agents that retrieve evidence and check hypotheses. A central challenge is distinguishing causal evidence from correlation: benchmark scores can hide weak evidence retrieval or unverified reasoning, and agents remain unreliable when telemetry is incomplete, misleading, or spread across long incident histories.

Latest papers

80 papers8 papers in the last four weeks, down 11% on the four weeks before. 0.0% of all new papers.

Every Alarm Has a Cause: An Open Generator of Telecom Alarm Storms with Exact Root Causes, Bound to NORIA-O, RFC 8632 and TMF642
Oct 21, 2026Madhulatha Mandarapu, Sandeep KunkunuruRoot Cause Analysis Fault Localization
VaidhyaMegha Private Limited, India
Semantic Observability for Autonomous Systems Diagnostics
Oct 20, 2026Jitendra Solanki, Anand Todkar, Ziran MinRoot Cause Analysis Knowledge Graphs
Siemens Corporation, Princeton, NJ, USA
Preventing Data Hallucinations in Agentic AI for Optical Network Troubleshooting
Oct 16, 2026Jeremias Dötterl, Martin KuipersHallucination Mitigation AI Agent Reliability
Adtran GmbH, Hermann-Dorner-Allee 91, 12489 Berlin, Germany · Adtran Networks SE, Hermann-Dorner-Allee 91, 12489 Berlin, Germany
Diviner: Agentic RCA with Human-Governed Adaptation for Large-Scale Infrastructure
Oct 15, 2026Pradipta Ghosh, Gaurav Mitra, Shripad Nadgowda +7Root Cause Analysis LLM Agent Reliability
Meta, Inc., USA
Agentic RCA for Internet-Scale Services Using Constrained Creativity
Oct 6, 2026Sayan Sinha, Vipul Harsh, B. Aditya Prakash +2Root Cause Analysis LLM Agents
Georgia Tech · Conviva · Carnegie Mellon University
SFlexRCA: Lightweight, Scalable, and Flexible Root Cause Analysis for IIoT Edge Systems
Oct 4, 2026Amr M. Zaki, Farhoud Jafari, Honggeun Ji +2Industrial Internet of Things Root Cause Analysis
Lassonde School of Engineering, York University, Toronto, ON, Canada
Beyond Conditional Independence: Root Cause Analysis with Deep Causal Models
Sep 29, 2026Md Musfiqur Rahman, Kenneth Lee, Ziwei Jiang +3Structural Causal Models Causal Inference
Electrical and Computer Engineering, Purdue University · Computer Science, Johns Hopkins University · Intuit AI Research +1
Where Root Cause Analysis Fails: A Retrieval-Reranking Decomposition
Sep 29, 2026Hada Melino Muhammad, Luan Pham, Laure Barrière +3Root Cause Analysis Industrial Anomaly Detection
University of New South Wales · Baker Hughes
Does Graph Structure Earn Its Place in Microservice Root-Cause Analysis? A Controlled Study on RCAEval, and What the Benchmark Was Really Measuring
Sep 22, 2026Imad BuljićGraph Neural Networks Benchmark Auditing
Polytechnic Faculty, University of Zenica
TraceVIC: Causal Reasoning over Code Evolution for Identifying Vulnerability-Inducing Commits
Sep 22, 2026Fnu Tanish, Samiha Shimmi, Samikshya Chapagain +3Software Vulnerability Detection Root Cause Analysis
Northern Illinois University · MIT Lincoln Laboratory
TriFleetRCA: On-Premise LLM Root Cause Analysis for Kubernetes
Sep 20, 2026Rohit Patel, Susil Kumar Mohanty, Jeenal ChaudharyLLM Inference Parallel LLM Inference
Department of Computer Science and Engineering Indian Institute of Technology Jodhpur, Jodhpur, India
From IceCube to IT-Sphere: A Hybrid Quantum-Classical GNN for Banking IT Root Cause Analysis
Sep 19, 2026Antonio Greco, Riccardo Paoletti, Roberto Cappuccio +1Root Cause Analysis Hybrid Quantum-Classical ML
Department of Physical Sciences, Earth and Environment (DSFTA), University of Siena, Siena, Italy · Istituto Nazionale di Fisica Nucleare (INFN), Pisa, Italy · IBM Consulting—Promontory, IBM Industry Diamond, Banking & Financial Markets, Milan, Italy
TriCalRAG: A Three-Strategy, Retrieval-Augmented Benchmark for On-Premise LLM-Based Root Cause Analysis in AIOps
Sep 13, 2026Rohit Patel, Susil Kumar Mohanty, Jeenal ChaudharyLLM Inference Efficiency Retrieval-Augmented Generation
Department of Computer Science and Engineering Indian Institute of Technology Jodhpur, Jodhpur, India
STAR: Adaptive Spatial-Temporal Normalization for Unified Microservice Incident Management
Sep 12, 2026Xinhua Miao, Linyu Zhu, Bowei Yang +1Self-Supervised Time Series Representation Learning Multimodal Anomaly Detection
School of Software Technology, Zhejiang University, Ningbo, China · School of Aeronautics and Astronautics, Zhejiang University, Hangzhou, China
TRACE: Training Reasoning Agents for Causal Exploration with Synthesized Rewards
Sep 9, 2026Rui Sun, Zhan Shi, Bing HeRL for Language Model Reasoning Causal Reasoning in Language Models
Independent Researchers
Large Language Models (LLMs) for Telecom Root Cause Analysis (RCA): A Structured Reasoning Framework for Evidence-Grounded Diagnosis
Sep 2, 2026Hao Zhou, Mandar Kulkarni, Hao Chen +3Root Cause Analysis LLM Reasoning
Jianzhong · Samsung Research America, Plano, Texas, TX 75023, USA
Loom: Weaving Diagnostic Strands into Free-Text Consensus via Embedding-Space Reweighting
Sep 2, 2026Ron Begleiter, Katya Egert Berg, Gilad Saban +1Root Cause Analysis
NVIDIA · Tel Aviv, Israel
ASTRA - Agentic System for Ticket Resolution and Analysis
Aug 28, 2026Shashidhar Reddy Javaji, Mohamed Trabelsi, Jin Cao +1Fault Diagnosis Evidence-Grounded Generation
Stevens Institute of Technology, Hoboken, New Jersey, USA · Nokia Bell Labs, Murray Hill, New Jersey, USA
TraceBench: Controlled Evaluation of LLM Agents for Time-Series Root-Cause Attribution
Aug 27, 2026Tommaso Bendinelli, Artur Dox, Christian HolzLLM Agent Evaluation Root Cause Analysis
Department of Computer Science ETH Zürich, Switzerland · CSEM SA, Switzerland · Independent Researcher
From General Agents to RCA Experts: A Self-Evolving Harness for Root Cause Analysis
Aug 26, 2026Haiyu Huang, Jiewei Lyu, Zhihan Jiang +5Root Cause Analysis LLM Agent Harnesses
The Chinese University of Hong Kong · Individual Researcher · ByteDance
Root cause analysis via difference graph discovery from linear time-series data
Aug 21, 2026Anouk Ruer, Timothée Loranchet, Daria Bystrova +1Root Cause Analysis Anomaly Localization
Sorbonne Université, INSERM, Institut Pierre Louis d’Epidémiologie et de Santé Publique, F75012, Paris, France
FinRCA-Bench: Benchmarking Evidence Retrieval and Reasoning for Financial AI Systems
Aug 19, 2026Pratik GhawateFinancial Reasoning Benchmarks Evidence Retrieval
Researcher
LongRCA Bench: Root-Cause Localization in Long-Horizon Agent Trajectories
Aug 15, 2026Yunfei Zhang, Boyu Feng, Changhua Pei +14Root Cause Analysis Agent Trajectory Analysis
Computer Network Information Center, Chinese Academy of Sciences · Chongqing University · Hangzhou Institute for Advanced Study, University of Chinese Academy of Sciences +4
The Abstention Protocol: RCA for Clos Fabrics
Aug 11, 2026Madhava Gaikwad, Deepak PandeyRoot Cause Analysis Cloud Computing
Independent · Microsoft
An Explainable GNN Framework for Component-Level Anomaly Diagnosis
Aug 10, 2026Sena Ozgunay, Louise Trav{é}-Massuy{è}s, Jean-Michel Loubes +1Industrial Anomaly Detection Root Cause Analysis
Comue de Toulouse, LAAS-DISCO, IMT · Comue de Toulouse, LAAS-DISCO · REGALIA, IMT +1
GALA: Graph-Augmented LLM Agents for Root Cause Analysis and Incident Response in Microservices
Aug 10, 2026Yifang Tian, Yaming Liu, Zichun Chong +3Tool-Augmented Language Model Agents LLM Agent Evaluation
University of Toronto Toronto, Canada
TelemetrySuffBench: Is Agent Telemetry Sufficient for Failure-Origin Diagnosis?
Aug 8, 2026Yuxuan Zhu, Peng PuSelective Prediction Fault Detection
School of Data Science and Engineering East China Normal University Shanghai, China
TELLER: Non-intrusive Cross-Layer Root-Cause Analysis for LLM Inference
Aug 3, 2026Ruilin Xu, Junyi Li, Pengfei Chen +1LLM Inference LLM Inference Serving
Sun Yat-sen University Guangzhou, China
ORCA-bench: How Ready Are Language Model Agents for Oncall?
Jul 30, 2026Albert Gong, Kyuseong Choi, Abhineet Agarwal +5Software Engineering Agents AI Agent Benchmarks
Cornell Tech · Traversal · Columbia University
EvoCause: LLM-Guided Evolution of Causal Graphs for Root Cause Analysis
Jul 29, 2026Lei Zan, Keli Zhang, Shifeng Xie +7Causal Discovery Root Cause Analysis
Huawei Noah’s Ark Lab · Guangdong University of Technology, Peng Cheng Laboratory

Oct 21, 2026J/K move · Enter open · S save

Every Alarm Has a Cause: An Open Generator of Telecom Alarm Storms with Exact Root Causes, Bound to NORIA-O, RFC 8632 and TMF642

Madhulatha Mandarapu, Sandeep Kunkunuru

VaidhyaMegha Private Limited, India

Root Cause Analysis Fault Localization

Open paper AskPDF arXiv

Alarm root-cause analysis is studied on data that cannot be shared. Operator logs are confidential, so what the field publishes is a causal graph between alarm types rather than per-incident causes, a fixed scenario set, gated, or never released. Two recent benchmarks carry per-incident root causes at scale; neither states a download location or a licence. We give a generator rather than a dataset. It runs on real backbone topologies from the Internet Topology Zoo and records the exact root cause and the propagation path of every alarm -- a by-product of generation, not a labelling pass. Every emitted field binds to a term in NORIA-O, RFC 8632, RFC 8345 or TMF642, and an unknown term fails the run. We ship fixed configurations, splits, a scorer and nine baselines. A one-line rule -- rank the device that raised the earliest critical alarm -- takes top-1 at 0.650 against 0.575 for our own temporal method, and wins 13 of 15 runs over three backbones and five seeds. An LLM given the same view reaches 0.600, the same model allowed to query the graph 0.625, and the rule wins on real operator alarms too, at 0.773: a property of alarm streams, not of our generator. One fault class separates the table: when a site loses power its PoP reports nothing, so every method that ranks alarmed devices scores 0.000 or near it, while the two that walk the topology to a silent candidate take 1.000 and 0.833.