logpai/logparser
Logparser is a machine learning toolkit that provides automated log parsing and benchmarks for structured log analytics via event template extraction.
Awesome AI for Infra › Log Analysis & Intelligence
logpai/Drain3
Drain3 is an open-source, robust Python library for online log template mining, offering critical functionality for AIOps and observability. It continuously processes a stream of raw log messages to identify and extract recurring patterns, transforming unstructured log data into structured templates. This process is essential for tasks like log anomaly detection, alert correlation, and root cause analysis. The project is an upgrade of the original Drain algorithm by LogPAI, enhanced for Python 3.6+ with new features including persistence (Kafka, Redis, file), streaming support, advanced masking of variable parts (e.g., IPs, numbers) to improve template accuracy, and efficient parameter extraction. Drain3 can operate in both training and inference modes, allowing for fast matching against learned templates without regular expressions. It includes configuration options for similarity thresholds, tree depth, and maximum cluster limits. By converting free-text log entries into structured templates, Drain3 significantly improves the utility of log data for AI-driven operational insights, making it a foundational component for intelligent monitoring and automation workflows in IT operations.
https://github.com/logpai/Drain3
Also in Log Analysis & Intelligence
logpai/loglizer
Loglizer is an open-source machine learning toolkit designed for automated anomaly detection in system logs, supporting various supervised and unsupervised models.
logpai/Log3C
Log3C is a framework that identifies impactful service system problems by analyzing system logs and KPI metrics through a process of log parsing, sequence vectorization, cascading clustering, and c...