The article explains how the ReAct reasoning‑acting loop transforms a generic AI troubleshooting agent into a precise fault‑diagnosis tool, boosting accuracy from about 30% to over 90% and even out‑performing human engineers, while providing a step‑by‑step Spring AI implementation.
Linux system administrators often encounter diverse failures, and this guide compiles over 40 distinct fault types—including system, network, hardware, and software issues—offering practical troubleshooting steps to help engineers quickly diagnose and resolve problems while building a solid knowledge base.
This article explains the hierarchical relationship between APM, distributed tracing, and observability, walks through a real Double‑11 e‑commerce incident, and demonstrates how a well‑designed observability stack can pinpoint the root cause, apply emergency fixes, and restore system performance within minutes.
Alibaba Cloud’s recent research breakthroughs—Mint, a cost‑efficient tracing framework that captures all request flows while drastically cutting storage and network overhead, and FAMOS, a multi‑modal fault‑diagnosis method for microservice systems—have been accepted to the prestigious ASPLOS and ICSE conferences, marking the first top‑conference publications in observability for the company.
ChatDBA is a conversational AI system built by Shanghai Aikesheng that employs large language models and Retrieval‑Augmented Generation to help database administrators diagnose faults, learn domain knowledge, and generate or optimize SQL, with a redesigned architecture that addresses early‑stage shortcomings and outlines future enhancements.
ChatDBA is an AI‑driven conversational system developed by Shanghai Aikesheng that assists DBAs with fault diagnosis, knowledge learning, SQL generation and optimization by leveraging large language models, RAG architecture, and advanced retrieval and document‑processing techniques.
ChatDBA, developed by Shanghai Aikesheng, is an AI-driven database operation assistant that leverages large language models and Retrieval‑Augmented Generation to provide fault diagnosis, knowledge learning, SQL generation and optimization, addressing challenges such as vague outputs, complex troubleshooting logic, and memory management through a structured architecture and multi‑modal retrieval strategies.
The article explains why collective communication is critical for distributed large‑model training, outlines the new requirements for system reliability, and introduces Baidu’s Collective Communication Library (BCCL), detailing its enhanced observability, fault‑diagnosis, stability, and performance optimizations that raise effective training time to 98 % and bandwidth utilization to 95 %.
Baidu’s Collective Communication Library (BCCL) enhances large‑model distributed training by improving real‑time bandwidth monitoring, fault diagnosis, network stability, and performance, leveraging RDMA networks and GPU‑specific optimizations to increase effective training time to 98% and bandwidth utilization to 95%.
This article explains the concept of distributed tracing, outlines its underlying principles in Java, details step‑by‑step implementation using popular SDKs, and describes common application scenarios such as performance monitoring, fault diagnosis, complex event handling, traffic analysis, and system optimization.
At KubeCon China 2023, experts Liu Kai and Dong Shandong presented a three‑part deep dive into Kubernetes observability challenges, demonstrating how eBPF enables comprehensive data collection across all stack layers, seamless integration, and intelligent root‑cause analysis through dimension attribution, anomaly bounding, and fault‑tree methods.
This talk explains the three major observability challenges in Kubernetes, demonstrates how eBPF enables comprehensive, low‑overhead data collection across all stack layers, and outlines a practical workflow that combines architecture awareness, application‑level metrics, and fault‑tree analysis to achieve automated root‑cause diagnosis.
This article details Guangdong Mobile's evolution of IT systems and operations, explains the four domain architecture, chronicles the AIOps adoption timeline, showcases intelligent anomaly detection, change assessment, fault diagnosis, and operation robots, and shares practical promotion methods and future outlook for AI‑driven IT operations.
This article details Ctrip's experience migrating MySQL workloads to the distributed OceanBase database, covering the design of an assessment tool, a one‑click migration workflow, comprehensive monitoring dashboards, automated fault‑diagnosis pipelines, encountered compatibility issues, and future roadmap for the platform.
This article introduces the concept of a data map, explains its background and goals, describes Youzan’s end‑to‑end data‑map practice—including full data lineage, search, management, link analysis, impact estimation, and optimization—and concludes with a summary and future outlook.
Xianyu’s production environment grew complex, prompting the creation of a high‑performance, sub‑3‑second real‑time data processing pipeline that ingests logs and metrics, uses Alibaba’s Logtail, LogHub, and Blink (enhanced Flink) for collection, transport, pre‑processing, computation, and persistent graph‑based fault analysis.
The article explains the March 23 Shanghai fiber‑optic outage that disrupted several Tencent apps and outlines eight typical reasons for fiber‑cable faults—such as construction cuts, vehicle accidents, fires, pole collisions, theft, animal damage, aging, and natural disasters—along with practical on‑site repair procedures.
This article explores Ctrip's adoption of AIOps—AI‑driven IT operations—detailing its concepts, typical use cases such as anomaly detection, intelligent fault diagnosis, and resource‑utilization improvement, and demonstrating how machine‑learning models like ARMA, FFT, and SVM have transformed operational efficiency, availability, and cost.
This article explores how rapid growth in network devices drives the need for comprehensive automation—covering script‑based tasks, zero‑touch provisioning, orchestration with OpenStack, device selection criteria, fault diagnosis, and monitoring—to keep operations ahead of business demands.
This article outlines a comprehensive approach for designing an IT operations monitoring platform that focuses on real‑time business transaction metrics, automatic topology discovery, event‑transaction correlation, deep component diagnostics, and unified data processing to improve availability, performance, and fault‑resolution speed in large‑scale data centers.
