LogAn - Large-scale Log Analysis with Small Language Models

Attend this 43-minute conference talk to discover LogAn, a revolutionary log analysis tool that leverages Small Large Language Models (SLMs) to transform IT troubleshooting and issue diagnosis. Learn from speakers Rahul Shetty and Siddardh R A as they demonstrate how this production-ready tool has been successfully deployed across 70 software products since March 2024, processing over 2000 tickets and saving 300+ man hours. Explore the current challenges in traditional IT log analysis methods and understand how LogAn addresses these obstacles by uncovering hidden insights that conventional analysis often misses. Discover techniques for optimizing LLM inference on CPU for processing large-scale log data efficiently, and understand how the tool performs insight extraction and causal analysis to summarize and identify root causes from high volumes of log data. Gain practical knowledge about implementing small language models for enterprise-level log analysis, specifically designed to empower Site Reliability Engineers (SREs) and Support Engineers. Experience a live demonstration showcasing how LogAn dramatically reduces incident detection and resolution time, and learn how to leverage this technology for root cause analysis in your own applications and troubleshooting workflows.