Missing-Aware Multimodal Fusion for Unified Microservice Incident Management

2026-03-26

AI Summary

This work addresses the challenge of missing multimodal monitoring data—metrics, logs, and traces—in microservice systems, which often arises from network or agent failures and degrades the performance of anomaly detection and root cause localization. Existing approaches rely on static placeholders that introduce noise and impair diagnostic accuracy. To overcome this limitation, the authors propose ARMOR, a self-supervised learning framework that jointly optimizes anomaly detection, fault classification, and root cause localization. ARMOR introduces two key innovations: modality-specific asymmetric encoders to decouple heterogeneous data distributions, and a missingness-aware gated fusion mechanism that combines learnable placeholders with dynamic bias compensation to effectively suppress cross-modal interference. Trained using only fault-type labels, ARMOR consistently outperforms state-of-the-art methods under both complete and severely incomplete data conditions, significantly enhancing the robustness and accuracy of microservice fault diagnosis.

Technology Category

• Machine Learning: Multimodal Learning
• Data Mining & Knowledge Management: Anomaly/Outlier Detection
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Graph Algorithms and Modeling for the Web: Representation, reconstruction, and subgraph or motif discovery in Web-related graphs

Abstract

Automated incident management is critical for microservice reliability. While recent unified frameworks leverage multimodal data for joint optimization, they unrealistically assume perfect data completeness. In practice, network fluctuations and agent failures frequently cause missing modalities. Existing approaches relying on static placeholders introduce imputation noise that masks anomalies and degrades performance. To address this, we propose ARMOR, a robust self-supervised framework designed for missing modality scenarios. ARMOR features: (i) a modality-specific asymmetric encoder that isolates distribution disparities among metrics, logs, and traces; and (ii) a missing-aware gated fusion mechanism utilizing learnable placeholders and dynamic bias compensation to prevent cross-modal interference from incomplete inputs. By employing self-supervised auto-regression with mask-guided reconstruction, ARMOR jointly optimizes anomaly detection (AD), failure triage (FT), and root cause localization (RCL). AD and RCL require no fault labels, while FT relies solely on failure-type annotations for the downstream classifier. Extensive experiments demonstrate that ARMOR achieves state-of-the-art performance under complete data conditions and maintains robust diagnostic accuracy even with severe modality loss.

Problem

• multimodal fusion
• missing modalities
• microservice incident management
• data incompleteness
• anomaly detection

Innovation

• missing-aware fusion
• multimodal self-supervised learning
• asymmetric encoder
• gated fusion mechanism
• microservice incident management