Tag
Loom presents a generative consensus framework for aggregating noisy textual hypotheses into consensus via embedding-space reweighting, deployed for real-world root cause analysis with improved efficiency and accuracy over traditional LLM agents.
This paper introduces CoreSec, a production system for root cause analysis in hyperscale datacenter networks that uses abstention algebra to manage ambiguous telemetry, ensuring deterministic and explainable results across Clos fabrics.
HookLens is a developer tool for real-time webhook triage and AI-powered root-cause analysis, aiding in issue diagnosis.
LongRCA Bench introduces a benchmark for diagnosing failures in long-horizon agent trajectories, and the RCTA method improves responsible role and root-cause step attribution.
Introduces ORCA-bench, a production-fidelity benchmark for evaluating LLM agents on oncall root cause analysis, finding that even frontier agents achieve only 25.3% accuracy on medium-difficulty tasks.
EvoCause is a research paper introducing an LLM-guided approach to refine causal graphs for root cause analysis, using expert diagnostic labels to constrain graph edits and releasing TeleRCA, an expert-annotated alarm benchmark from a production telecom network.
This paper proposes weighted conformal methods for changepoint localization and root cause analysis that reduce confidence set size under corrupted observations by downweighting likely contaminated data, using uncertainty signals and meta-learning.
This article draws parallels between reliability in manufacturing and modern software engineering, highlighting principles like redundancy, root cause analysis, and observability to build resilient systems.
Sentry released an AI agent called 'Seer' that analyzes backend errors, determines root cause, drafts a fix, and automatically opens a pull request for review.
This paper studies root cause analysis on real-world telemetry data using the OpenRCA benchmark, showing that existing classical and LLM-based methods fail and proposing a Structured Multi-Agent RCA pipeline that substantially outperforms them. It further reveals through reverse reasoning that the primary bottleneck is reasoning capability rather than data access, and introduces automated rule mining to reduce reliance on manual domain knowledge.
This paper audits offline root-cause-analysis benchmarks and finds that pooled leaderboards hide subsystem-specific winners, using pairwise comparisons on 778 cases across 11 subsystems. It releases a 320-line audit module for recomputing per-subsystem stability checks.
StableRCA is a novel root cause analysis framework that identifies intervention targets by estimating local Markov boundaries and detecting conditional distribution shifts, avoiding the need for global causal graph discovery and demonstrating robustness across synthetic and real-world datasets.
This paper introduces a formal definition of causal pathways for rare events and discusses testable implications, bridging simple verbal explanations with detailed causal models.
ORCA is a copilot for end-to-end causal analysis that uses agents to guide users through workflows including causal discovery, effect estimation, and root cause analysis, with structured reports.
TopoEvo is a topology-aware self-evolving multi-agent framework for root cause analysis in microservices that couples graph representation learning with structured, topology-constrained reasoning. It achieves absolute improvements of up to 3.44% in root cause localization accuracy and boosts fault-type classification performance by 4.39% to 16.81% across diverse datasets.
STAR is a stage-attributed triage and repair framework that decomposes LLM-based RCA agent workflows into four structured stages, enabling stage-wise auditing, counterfactual evaluation, and patch-and-replay repair to improve root cause localization and fault type classification in microservice AIOps.
This guide from OpenAI Academy explains how data science teams can use Codex to speed up analysis workflows, including root-cause analysis, business impact readouts, and handling ambiguous requests.