Bridging Temporal and Textual Modalities: A Multimodal Framework for Automated Cloud Failure Root Cause Analysis

TL;DR

This paper introduces a multimodal framework that combines time-series data and language models to improve automated root cause analysis in cloud failures, achieving high diagnostic accuracy.

Contribution

It proposes a novel alignment method and semantic compression technique to enable language models to interpret continuous temporal data for incident diagnosis.

Findings

Achieves 48.75% diagnostic accuracy on cloud benchmarks.

Outperforms existing methods in compound failure scenarios.

Demonstrates effective multimodal data integration for incident analysis.

Abstract

Root cause analysis in modern cloud infrastructure demands sophisticated understanding of heterogeneous data sources, particularly time-series performance metrics that involve core failure signatures. While large language models demonstrate remarkable capabilities in textual reasoning, their discrete token-based architecture creates fundamental incompatibilities with continuous numerical sequences exhibiting temporal dependencies. Current methodologies inadequately address this modality mismatch, constraining the potential of language model-driven automation in incident management workflows. This paper presents a multimodal diagnostic framework that harmonizes time-series representations with pretrained language model embedding spaces. Our approach contributes three technical advances: (1) a semantic compression technique that distills temporal segments into single-token abstractions while preserving pattern semantics, (2) an alignment encoder utilizing gated cross-attention to project time-series features into language model latent space, and (3) a retrieval-augmented diagnostic pipeline that synthesizes aligned embeddings with historical incident knowledge for expert-level failure attribution. Comprehensive evaluation across six cloud system benchmarks demonstrates that our framework achieves leading performance, reaching 48.75% diagnostic accuracy with notable improvements on scenarios involving compound failure modes. The results validate embedding-space alignment as an effective strategy for enabling language models to reason over multimodal telemetry data in production incident response contexts.