Nezha: Interpretable Fine-Grained Root Causes Analysis for Microservices on Multi-modal Observability Data
The DOI matches its Crossref record; 46 of the 46 references it cites were found in a registry or an index.
Abstract
Root cause analysis (RCA) in large-scale microservice systems is a critical and challenging task. To understand and localize root causes of unexpected faults, modern observability tools collect and preserve multi-modal observability data, including metrics, traces, and logs. Since system faults may manifest as anomalies in different data sources, existing RCA approaches that rely on single-modal data are constrained in the granularity and interpretability of root causes. In this study, we present Nezha, an interpretable and fine-grained RCA approach that pinpoints root causes at the code region and resource type level by incorporative analysis of multi-modal data. Nezha transforms heterogeneous multi-modal data into a homogeneous event representation and extracts event patterns by constructing and mining event graphs. The core idea of Nezha is to compare event patterns in the fault-free phase with those in the fault-suffering phase to localize root causes in an interpretable way. Practical implementation and experimental evaluations on two microservice applications show that Nezha achieves a high top1 accuracy (89.77%) on average at the code region and resource type level and outperforms state-of-the-art approaches by a large margin. Two ablation studies further confirm the contributions of incorporating multi-modal data.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Nezha [37] jointly encodes metrics, logs, and traces into a shared space with contrastive alignment and performs graph-based reasoning to localize causally relevant services under noisy/partial observability.”
From TopoEvo: A Topology-Aware Self-Evolving Multi-Agent Framework for Root Cause Analysis in Microservices (2026) · cites this paper as [37]
Citations
Cites
USAD: UnSupervised Anomaly Detection on Multivariate Time Series
Audibert, J., Michiardi, P., Guyard, F., Marti, S., & Zuluaga, M. A. (2020) · Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining · doi:10.1145/3394486.3403392 · Cited by 1,094 · cited 4×
“, USAD [2] and JumpStarter [43]) and monitor additional metrics beyond the P90 latency and success ratio to avoid missing faults.”
BackgroundInfluentialDrain: an online log parsing approach with fixed depth tree
He, P., Zhu, J., Zheng, Z., & Lyu, M. R. (2017) · 2017 IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws.2017.13 · Cited by 943 · cited 2×
“Data Integrator adopts a state-of-the-art log parsing approach Drain [19] to extract the static log templates and dynamic log parameters from the raw log messages in a streaming manner.”
MethodSelf-citationVisualization of test information to assist fault localization
Jones, J. A., Harrold, M. J., & Stasko, J. (2002) · Proceedings/Proceedings - International Conference on Software Engineering · doi:10.1145/581339.581397 · Cited by 904
Log clustering based problem identification for online service systems
Lin, Q., Zhang, H., Lou, J.-G., Zhang, Y., & Chen, X. (2016) · Proceedings of the 38th International Conference on Software Engineering Companion · doi:10.1145/2889160.2889232 · Cited by 427
Zhou, X., Peng, X., Xie, T., Sun, J., Ji, C., Li, W., & Ding, D. (2021) · IEEE Transactions on Software Engineering · doi:10.1109/tse.2018.2887384 · Cited by 410 · cited 3×
“The complexity of fault scenarios is alleviated by injecting resource issues and code defects from real faults in industrial systems [33, 76] into microservices.”
BackgroundResultMicroRCA: Root Cause Localization of Performance Issues in Microservices
Wu, L., Tordsson, J., Elmroth, E., & Kao, O. (2020) · NOMS 2020 - 2020 IEEE/IFIP Network Operations and Management Symposium · doi:10.1109/noms47738.2020.9110353 · Cited by 265
Unsupervised Detection of Microservice Trace Anomalies through Service-Level Deep Bayesian Networks
Liu, P., Xu, H., Ouyang, Q., Jiao, R., Chen, Z., Zhang, S., Yang, J., Mo, L., Zeng, J., Xue, W., et al. (2020) · 2020 IEEE 31st International Symposium on Software Reliability Engineering (ISSRE) · doi:10.1109/issre5003.2020.00014 · Cited by 252 · cited 5×
“We use fault injection to mimic application issues following the previous work [35, 37, 42, 68, 72, 73].”
MethodResultInfluentialDeepTraLog: trace-log combined microservice anomaly detection through graph-based deep learning
Zhang, C., Peng, X., Sha, C., Zhang, K., Fu, Z., Wu, X., Lin, Q., & Zhang, D. (2022) · Proceedings of the 44th International Conference on Software Engineering · doi:10.1145/3510003.3510180 · Cited by 207
Root Cause Analysis: A Tool for Total Quality Management
Wilson, P. F., Dell, L. D., & Anderson, G. F. (1996) · Journal For Healthcare Quality · doi:10.1111/j.1945-1474.1996.tb00823.x · Cited by 204
Identifying impactful service system problems via log analysis
He, S., Lin, Q., Lou, J.-G., Zhang, H., Lyu, M. R., & Zhang, D. (2018) · Proceedings of the 2018 26th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering · doi:10.1145/3236024.3236083 · Cited by 201
Research works printed but not linked: 8
- Michael Chow, David Meisner, Jason Flinn, Daniel Peek, and Thomas F. Wenisch. 2014. The Mystery Machine: End-to-end Performance Analysis of Large-scale Internet Services. In OSDI 2014. USENIX Association, 217–231.
- Rodrigo Fonseca, George Porter, Randy H. Katz, Scott Shenker, and Ion Stoica. 2007. X-Trace: A Pervasive Network Tracing Framework. In NSDI, 2007. USENIX, 271–284.
- Chang Lou, Peng Huang, and Scott Smith. 2020. Understanding, Detecting and Localizing Partial Failures in Large System Software. In NSDI 2020. USENIX Association, 559–574.
- Minghua Ma, Shenglin Zhang, Junjie Chen, Jim Xu, Haozhe Li, Yongliang Lin, Xiaohui Nie, Bo Zhou, Yong Wang, and Dan Pei. 2021. Jump-Starting Multivariate Time Series Anomaly Detection for Online Service Systems. In USENIX ATC 2021. USENIX Association, 413–426.
- Karthik Nagaraj, Charles Edwin Killian, and Jennifer Neville. 2012. Structured Comparative Analysis of Systems Logs to Diagnose Performance Problems. In NSDI 2012. USENIX Association, 353–366.
- Opentelemetry. 2023. Opentelemetry. https://opentelemetry.io Accessed Jan. 6, 2023
- Benjamin H Sigelman, Luiz Andre Barroso, Mike Burrows, Pat Stephenson, Manoj Plakal, Donald Beaver, Saul Jaspan, and Chandan Shanbhag. 2010. Dapper, a large-scale distributed systems tracing infrastructure.
- Apache SkyWalking. 2023. Apache SkyWalking. https://skywalking.apache.org Accessed Jan. 6, 2023
Cited by
Items 101–139 of 139
Root Cause Analysis for Middleware Issues by Kubernetes Resource Events
Koyama, T., Kushida, T., & Ikuno, S. (2026) · International Conference on Knowledge and Smart Technology · doi:10.1109/kst67832.2026.11431846 · Cited by 1
Xu, Y., & He, G. (2026) · International Conference on Artificial Intelligence and Social Network Systems (AISNS) · doi:10.1109/aisns67921.2026.11440435
“For AD, methods leverage logs and metrics to detect abnormal behaviors [2]–[4]; for FT, solutions categorize incidents based on observed patterns [5], [6]; and for RCL, candidate instances are ranked by correlating signals with system-level anomalies [7]–[10].”
Adaptive Anomaly Detection for Multi-Dimensional Issue Reports
Wang, X., Chu, M., Yu, H., Luo, C. W., & Hu, C. (2026) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3801155
Token interdependency parsing (Tipping)—A map-reduced based statistical log parser
Hashemi, S., & Mäntylä, M. (2026) · Information and Software Technology · doi:10.1016/j.infsof.2026.108137 · Cited by 1
Liu, M., Li, Y., Wang, H., Li, M., Tang, F., & Yang, W. (2026) · IEEE Transactions on Reliability · doi:10.1109/tr.2026.3680212
“To solve this problem, some works input timestamps as text into the language encoding network to align temporal information [10], [36].”
Self-citationA tolerance analysis framework for microservice-based systems against cascading failures
Zheng, C., Si, S., Wang, X., Wang, J., Zhang, X. X., Lv, S., & Sun, L. (2026) · Cybersecurity · doi:10.1186/s42400-025-00399-w · Cited by 1
Representation-Aware Root Cause Analysis with Large Language Models (Position Paper)
Wen, Y., Panahandeh, M., Chouchen, M., & Hamou‐Lhadj, A. (2026) · ACM/SPEC International Conference on Performance Engineering (ICPE) · doi:10.1145/3777911.3801108
“A large body of prior work addresses RCA in microservices by analyzing observability data to identify anomaly propagation and fault localization patterns [8, 9, 19].”
Tian, X., Ying, S., Li, T., Li, N., Shi, C., & Xiao, D. (2026) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2026.3687558
“In this section, we introduce three motivations with an example from a widely-used benchmark microservice system TrainTicket [1], [2], [5], [6], [26].”
InfluentialUrbanke, P., & Fischer, S. (2026) · IEEE International Conference on Software Analysis, Evolution and Reengineering - Companion (SANER-C) · doi:10.1109/saner-c67878.2026.00051
“Several datasets have been published for the TrainTicket system, including [19]–[22], but they exhibit limitations that reduce their usability for fault analysis.”
STAR: A Stage-Attributed Triage and Repair Framework for RCA Agents in Microservices
Wang, J., Chen, R., Wang, Y., Liao, X., Fan, Z., Wu, X., Wu, W., Zhao, J., & Liu, X. (2026) · IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws72778.2026.00057 · Cited by 4
“Yet such systems remain vulnerable to reasoning failures such as evidence omission, hypothesis drift, and decision inconsistency, while lacking an explicit mechanism to localize and repair the faulty reasoning stage.”
Topology-Aware Circuit Breaking on Critical Paths in Microservice Systems
Wang, L., Li, X., Bu, Y., Zhang, T., Teng, M., & Zhao, Y. (2026) · Design, Automation, and Test in Europe · doi:10.23919/date69613.2026.11539340
Self-citationToward In-Depth Root Cause Localization for Microservices With Multi-Agent Recursion-of-Thought
Zhang, L., Jia, T., Wang, K., Duan, C., He, M., Wang, R., Peng, X., Wang, M., Zhang, G., Chen, R., et al. (2026) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2026.3706572 · Cited by 4
“Specifically, compared with the strongest LLM-based baseline in terms of speed, namely ReAct, RCLAgent reduces the average inference time from 87.63 s to 41.43 s on AIOPS 2022, from 89.73 s to 59.55 s on TrainTicket, and from 85.77 s to 49.79 s on RCAEval.”
InfluentialSelf-citationPRISM: Cross-Layer Observability Framework for Mobile Core Networks
Samarakoon, S., Mohan, N., & Kuipers, F. A. (2026) · IFIP Networking Conference (IFIP Networking) · doi:10.23919/ifipnetworking70592.2026.11579127
“Among them, graph-based RCA approaches go beyond multi-modal data collection by correlating heterogeneous operational signals, such as logs, metrics, and traces, through structural dependency models [17]–[20].”
Gao, Y., Yang, L., Meng, L., Zhu, R., & Cao, Y.-N. (2026) · Proceedings of the ACM on software engineering. · doi:10.1145/3808180
ProfRCA: LLM-Enabled Fine-Grained Root Cause Analysis with Continuous Profiling Data
Ye, S., Tan, G., Yang, W., & Chen, P. (2026) · IEEE International Conference on Software Analysis, Evolution and Reengineering (SANER) · doi:10.1109/saner67736.2026.00111 · Cited by 2
“Traces track end-to-end requests across services to reveal performance bottlenecks [33], [48].”
Self-citationAI-NativeBench: An Open-Source White-Box Agentic Benchmark Suite for AI-Native Systems
Wang, Z., Yu, G., & Lyu, M. R.-T. (2026) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3821576 · Cited by 6
Self-citationLeveraging Hyperbolic Geometry for Enhanced Graph-based Anomaly Detection in Microservices
Wu, Z. W., Han, Y., Zhang, W., & Wei, W. (2026) · International Conference on Computer Supported Cooperative Work in Design (CSCWD) · doi:10.1109/cscwd68734.2026.11582829
Ding, K., Ran, D., Ma, Y., Song, Z., Bian, K., & Xie, T. (2026) · IEEE Transactions on Dependable and Secure Computing · doi:10.1109/tdsc.2026.3712748
“Current approaches primarily leverage three data sources: metrics, logs, and traces, which collectively provide comprehensive insights into microservice behavior [1], [3].”
Harnessing LLMs with Graph Representation Learning for Anomaly Detection in Microservices
Wu, Z., Han, Y., & Wei, W. (2026) · International Conference on Communications · doi:10.1109/icc59461.2026.11587392
“[3] and [4] utilized a graph to jointly model trace and log entries.”
Uncovering Microservice Faults: A Temporal Graph Approach to Root Cause Analysis
Aharon, U., Dvir, A. Z., Dubin, R., Marbel, R., & Hajaj, C. (2026) · International Conference on Communications · doi:10.1109/icc59461.2026.11586787
“Yu et al. [11] rank event patterns by their frequency.”
Anomaly Detection in Microservices Using Word2Vec with Bidirectional Long Short-Term Memory Method
Rasheedh, J. A., Zafer Ahmed, N., Ahmed, G. N., & Hakkim, S. (2025) · International Conference on I-SMAC (IoT in Social, Mobile, Analytics and Cloud) (I-SMAC) · doi:10.1109/i-smac65734.2025.11597592
MicroInspect: A Progressive Framework for Metric-Level Root Cause Localization in Microservices
Yin, Y., Zhu, T., Chen, T., Lv, M., Song, Q., & Ji, B. (2026) · IEEE Transactions on Services Computing · doi:10.1109/tsc.2026.3713967
Freezing the Crime Scene: A State Snapshot Paradigm for Reproducible Agentic SRE Evaluation
Yu, G., Wang, Y., & Lyu, M. R.-T. (2026) · ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE) · doi:10.1145/3803437.3806092
Self-citationAIMS: A Content-Aware Resource Management Approach for AI Assistant Systems
Duan, C., Jia, T., He, M., Xiao, P., Zhang, L., Zhong, Z., Zhang, X., & Yang, Y. (2026) · ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE) · doi:10.1145/3803437.3805562 · Cited by 1
“While recent work on microservice performance debugging [5, 7] and root cause analysis [6, 21] has improved our understanding of system failures, these approaches primarily focus on post-failure diagnosis rather than proactive prevention.”
Ding, K., Song, Z., & Bian, K. (2026) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3770854.3780304 · Cited by 1
“Independent modeling approaches [17, 34, 37] enhance noise robustness but lose cross-variable dependencies.”
InfluentialTSLoc: Self-Supervised Faulty Node Localization Framework in Large-scale Training Clusters
Zhou, Q., Pei, C., Lu, Y., Ma, D., Wang, Z., Li, J., Zhu, Y., Jiang, D., Pei, D., Li, J., et al. (2026) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3770855.3818499
TD-RCA: Topology-Aware and Dual-Perspective Decoupling for Root Cause Analysis in Microservices
He, S., Tang, B., Wei, M., Xiang, L., & Xie, K. (2026) · International Workshop on Quality of Service · doi:10.1109/iwqos70441.2026.11661074
“Nezha [28] and MULAN [29] introduce multi-modal fusion, but missing trace ID s cause alignment failure.”
ChatRCA: A Root Cause Analysis Method via LLMs-based Multi-Agent with Human-in-the-Loop
Hui, M., Wang, L., Li, Q., Zhang, L., Bai, Z., Hu, J., Li, H., Ren, Y., Song, F., Wang, Y., et al. (2026) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3842744 · Cited by 1
“We evaluate ChatRCA on three datasets: the open-source TrainTicket dataset [44], a private dataset collected from real cloud incidents in an enterprise cloud platform, and the open-source GAIA dataset.”
Reliable Orchestration of Specialized Multi-Agents for Cloud Incident Diagnosis
Huang, H., Lyu, J., Lai, S., Jiang, Z., Chen, P., & Lyu, M. R.-T. (2026) · IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws72778.2026.00187
“Although each pillar is individually valuable, their heterogeneous formats and differing noise characteristics make joint analysis non-trivial [2], [6].”
InfluentialSelf-citationSCALER: LLM-Based Cross-Modal Alignment for Microservice Root Cause Analysis
Yang, S., Wang, G., Ming, Z., He, D., Cui, L., Yu, X., & Zhou, X. (2026) · IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws72778.2026.00101
“Representative works, such as Nezha [40], MULAN [10], HeMiRCA [12], and RCRank [13], further demonstrate the effectiveness of multi-modal alignment and causal modeling in microservice systems.”
DCRL: Dynamic Causal Representation Learning for Root Cause Localization in Microservice Systems
Liu, C., Zhang, J., Shi, Y., Liu, S., Zhang, D., Qi, K., & Chen, Z. (2026) · IEEE International Conference on Web Services (ICWS) · doi:10.1109/icws72778.2026.00080
“With the rapid advancement of deep learning, an increasing number of studies have leveraged graph representation learning to address root cause localization in microservice systems [22]–[27].”
A Lightweight Causal Evidence Pipeline for Root-Cause Localization in Cloud-Native Systems
Dua, M., Gautam, A. S., Chowdhury, R., & Kundu, T. (2026) · IEEE International Conference on Information Reuse and Integration (IRI) · doi:10.1109/iri69576.2026.00015
“A direct comparison with full systems (CloudRanger [4], TORAI [5], Nezha [9], MicroRCA [11]) requires standardized benchmarks such as RCAEval [14] and is planned as immediate future work.”
FoundRoot: Towards Foundation Model for Root Cause Analysis via Structured Deep Thinking
Xie, Z., Li, Z., He, X., Zhang, S., Xu, L., Yang, Y., Zhang, T., Chen, J., Shi, R., & Pei, D. (2026) · ACM/IEEE International Conference on Software Engineering: Companion Proceedings (ICSE-Companion) · doi:10.1145/3744916.3787814 · Cited by 7
Scalable and Adaptive Log-based Anomaly Detection: A Synergistic Approach
Liu, J., Huang, J., Jiang, Z., Huo, Y., Chen, Z., Gu, J., Feng, C. Q., Yan, M., & Lyu, M. R.-T. (2026) · ACM Transactions on Software Engineering and Methodology · doi:10.1145/3846182
Gleaner: A Semantically-Rich and Efficient Online Sampler for Microservice Diagnostics
Yang, Y., Fang, A., Zhang, S., & He, P. (2026) · Proceedings of the ACM on software engineering. · doi:10.1145/3832241
“We evaluate how sampler choice affects automated Root Cause Analysis (RCA) accuracy using three open-source RCA algorithms: MicroRCA [38], Nezha [44], and ShapleyIQ [25].”
InfluentialOpsLens: Runbook-Free Root Cause Analysis via Multi-agent Collaboration and Tool Management
Huang, H., Lyu, J., Tan, G., Chen, P., Jiang, Z., Yu, G., Mi, P., Wang, F., Jian, H., Tan, C., et al. (2026) · IEEE/ACM International Conference on Automated Software Engineering (ASE) · doi:10.1145/3832783.3834501
“RootCauseAnalysis(RCA)playsakeyroleinmaintainingcloud reliability[22,38,45,46].”
InfluentialSelf-citationHERO: Hypothesis-Centered Root-Cause Analysis for Microservice Incidents
Lyu, J., Yi, J., Liang, S., Chen, P., & Pan, L. (2026) · IEEE/ACM International Conference on Automated Software Engineering (ASE) · doi:10.1145/3832783.3837509
“[44] A multi-modal approach that represents metrics, traces, and logs as unified events and mines event-graph patterns to localize fine-grained root causes.”
Self-citationAlarmClaw: Context-Enriched Alarm Management with Category-/Severity-Aware Incident Graphs
Yu, S., Li, M., Yang, J., Wu, Y., Wang, N., Chen, Z., Liu, M., 任心池, Ren, X., Li, C., et al. (2026) · IEEE/ACM International Conference on Automated Software Engineering (ASE) · doi:10.1145/3832783.3834524
Self-citationEvaluating Diagnostic Evidence in Multimodal Network Fault Classification
LEE, J., Oh, S., Park, J., & Kim, J. (2026) · Processes · doi:10.3390/pr14203341
Metadata
- Authors
- Yu, G., Chen, P., Li, Y., Chen, H., Li, X., & Zheng, Z.
- Affiliations
- Sun Yat-sen University (CN)
- Venue
- Proceedings of the 31st ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering
- Publisher
- ACM
- Published
- 2023-11-30
- Pages
- 553-565
- OpenAlex
- W4389158453
- Cited by
- 140
- References
- 46
- Citations by year
- 2026: 59 · 2025: 58 · 2024: 23
- FWCI
- 42.18
- Type
- conference paper
- Language
- en
- Open access
- closed
- Locations
- ACM Joint European Software Engineering Conference and Symposium on the Foundations of Software Engineering (ESEC/FSE) (publishedVersion)
- Funders
- National Key Research and Development Program of China (2019YFB1804002)
- National Natural Science Foundation of China (No.62272495)
- Guangdong Basic and Applied Basic Research Foundation (No.2023B1515020054)
- CCF-Lenovo Blue Ocean Research Fund
- Topics
- Physical Sciences › Computer Science › Computer Networks and Communications › Software System Performance and Reliability
- Physical Sciences › Computer Science › Software › Software Testing and Debugging Techniques
- Physical Sciences › Computer Science › Information Systems › Software Engineering Research
- Keywords
- root cause analysis, microservices, event graphs, fault location, interpretable fault diagnosis
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.